High current H-Bridge? Greater than 10A @ 12V
eagletalontim
Posts: 1,399
in Propeller 1
One of these days I will feel more comfortable selecting new components on my own, but for now, I would like some opinions on something I am working on.
My project is the same as my solar panel tracker I have already completed but I have stepped up and have 10 panels to move all at the same time on a tilting frame. I have selected a 24" 1000lb rated linear actuator that is 12V and has an amp draw of 10A at full load. No load is 3A. My current component is this : http://www.digikey.com/product-detail/en/0/TLE5206-2S-ND It is only rated for 5A which is just not going to cut it. So...I have been searching for the next best component that will fit on my circuit board and the Prop can drive it. No PWM is needed. The longest one of the actuators will be on will be a max of 20 to 30 seconds when the system "wakes" up in the morning and when it "parks" at night or if the wind picks up too much.
I am trying my best to stay away from relays since they are mechanical. I have limited room on my circuit board so I need to try and keep the surface area as small as possible. The 220-7 pin fits quite nicely and I may be able to squeeze in a little more if needed. I will use the same circuit for the other linear actuator as well.
Any help is much appreciated!
My project is the same as my solar panel tracker I have already completed but I have stepped up and have 10 panels to move all at the same time on a tilting frame. I have selected a 24" 1000lb rated linear actuator that is 12V and has an amp draw of 10A at full load. No load is 3A. My current component is this : http://www.digikey.com/product-detail/en/0/TLE5206-2S-ND It is only rated for 5A which is just not going to cut it. So...I have been searching for the next best component that will fit on my circuit board and the Prop can drive it. No PWM is needed. The longest one of the actuators will be on will be a max of 20 to 30 seconds when the system "wakes" up in the morning and when it "parks" at night or if the wind picks up too much.
I am trying my best to stay away from relays since they are mechanical. I have limited room on my circuit board so I need to try and keep the surface area as small as possible. The 220-7 pin fits quite nicely and I may be able to squeeze in a little more if needed. I will use the same circuit for the other linear actuator as well.
Any help is much appreciated!
Comments
Sandy
For higher currents, you may be better selecting a Driver + FETS,
This part, recently announced appealed to me.
http://www.ti.com/product/DRV8701
It has wide Vcc 5.9-V to 45-V and can do 100% drive (includes Charge pumps) and includes modest LDOs
Just add four external N-channel MOSFETs, of AMPS to suit.
My circuit uses an XBee to communicate with my website for data feedback and controlling via a web browser and putting something that could cause electrical noise that close to my XBee would probably kill the reliability.
If something is cheap or not depends on what criteria you apply. My application uses 6 of the CMX100D10 relays and I consider them dirt cheap because of their reliability and the equipment downtime they save.
Reliable operation equals happy customer.
I am open to ideas but as stated earlier, I would prefer not to use relays for this application. Solid states are also way too expensive for my taste :P
Another approach is to use something like Infineon Smart High Side Switch | PROFET™
http://www.infineon.com/cms/en/product/parametricSearch.html?channel=ff80808112ab681d0112ab69e2d40357&showAllParameters=true&sort=group&sortDir=&parentChannelRef=&ipt=false&5546d4694909da48014909dc03640175_true=High-Current+PROFET&intc=0120097
and use generic N-FETS on the low side.
The component I listed HERE seems to be 1 component for each motor, looks like it would only need a couple of resistors between it and the Prop, and can handle everything I need all in one chip. Less components to work with, less I could mess up :P Question is, can the Prop drive it without using extra components besides a couple resistors?
That has 2 FETS, so you need 2 for a full H bridge.
Another approach is to use 2 x N-FETS and a SPCO relay.The SPCO is changed when N-FETS are off, so never makes or breaks current, & the FETs control the Motors, and can soft-start as needed.With some care, the relay should change only twice a day ?
BJT transistors will create a great deal of heat and a big voltage drop. You would have to at least provide 14VDC to get 12VDC to the motor and more likely 16VDC. The heat would mean large heat sinks.
So for solid-state, MOSfet seems to be best. And the MOSfet solutions really tend to be mostly available in fabricated boards. I suppose one might trying building their own MOSfet solution from scratch, but devices that are better are mainly surface mount.
BJT transistors will create a great deal of heat and a big voltage drop. You would have to at least provide 14VDC to get 12VDC to the motor and more likely 16VDC. The heat would mean large heat sinks.
So for solid-state, MOSfet seems to be best. And the MOSfet solutions really tend to be mostly available in fabricated boards. I suppose one might trying building their own MOSfet solution from scratch, but devices that are better are mainly surface mount.
Not sure at all how you came to that conclusion as you may be thinking of darlingtons which I would not use for continuous power applications due to the high Vce(sat). There are plenty of BJTs that exhibit very low Vce(sat) of maybe 200mv with high Hfe current gain that would work well although in saying this I would still opt for a MOSFET.
The TIP122s and such like even ULN2803s that seem to be specified by many fritzing Adunio aficionadoes are fine for low (very low) duty cycle use but are not meant for continuous current due to the high power dissipation etc. But these devices do have their place nonetheless, rarely where they seem to be used though.
@eagletalontim: You know the relay + MOSFET idea mentioned is not really so bad an idea for your application, it keeps it simple and the relay never has to worry about switching high currents, just switching and then the high currents are turned on by the rugged N-channel MOSFETs (I have a few in mind). Relays can become a problem when the contacts arc and spark due to switching high currents, not something that they are required to do in this app. The fact that you only have low-side N-channel MOSFETs which are easy to drive from the Prop keeps everything simple and hopefully foolproof.
10 AMPs.... The ULN2803 is a Darlington with incumbent voltage drops and only 60ma per channel.
And the TIP122 is a Darlington as well, but only half an H-bridge.
Usually when you jump on me your a spot on, but it seems you are off your game today.
@ Peter
10 AMPs.... The ULN2803 is a Darlington with incumbent voltage drops and only 60ma per channel.
And the TIP122 is a Darlington as well.
Usually when you jump on me your a spot on, but it seems you are off your game today.
hehe, you think I jump on you, do you? But I only mentioned 2803s because in a similiar manner they seem to be used because they are rated at 800ma "each".....not (RTM).
DING.... round 2
The TIP122 is only rated up to 5 amps. Yet another miss.
Not at all, I don't think there is a darlington suitable for 10A+, how could there be? If typical Vce(sat) at that kind of current could be well over 1V then 10W+ is ridiculous. Rather though, my quoting of part numbers is not in any shape or form a "recommendation" but simply a reference to a couple of well known examples of that kind and their unsuitability in most applications to which fritzingers apply them to.
But we have to come full circle back to your statement which seems to infer that your knowledge of power transistors is probably in similar manner darlington centric because of your statement "BJT transistors will create a great deal of heat and a big voltage
drop. You would have to at least provide 14VDC to get 12VDC to the
motor and more likely 16VDC. The heat would mean large heat sinks."
I rest my case.
I guess I didn't read your original reply correctly.
You claim there are some useful BJTs for 10amp, 12VDC -- but don't specify anything. Maybe there are, but I don't have any idea what they are.
Besides, MOSfets can be put in parallel - unlike the BJTs. That makes it easier to get high amperage in available devices.
What makes you think though that BJTs can't be in parallel? I think it is this "secret knowledge" you flout that flies contrary to fact that I contest indirectly simply to set matters straight.
What makes you think though that BJTs can't be in parallel? I think it is this "secret knowledge" you flout that flies contrary to fact that I contest indirectly simply to set matters straight.
The audio world attempted BJTs in parallel and had problems with thermal runaway. It may be 'myth', but certainly not secret knowledge. Yes indeed, your indirectness is confusing. But confusion reigns these daze.
On the other hand, I have seen lots of MOSfets used in parrallel, mostly in electric bike battery chargers.
Good Morning
Hi,
You selected a Pololu product.. which is what I would have gone with as they design well and work hard to use some pretty leading edge MOSfet H-bridges.
Building would have been a big headache.
Pololu says this -- "The logic connections are designed to interface with 5V systems (5.5 V
max); the minimum high input signal threshold is 3.5 V, so we do not
recommend connecting this device directly to a 3.3 V controller."
And it seems to require a separate +5VDC power to the internal logic.
You could just buy a tiny 3.3VDC to 5VDC level shifter board from SparkFun or EBay.
You just add some small resistances in series with each emitter if you connect power BJTs
in parallel - this prevents current-hogging by the hotest device. You get some benefit from
mounting them close together on the heatsink too.
I could not see where it needs a separate 5V power supply. On the logic side of the board, it shows a 5V output that can supply a few mA.
Or you being the shopper could do the same.
And there might be something more suitable for 3.3V logic at Pololu. I haven't shopped all the specifics. I just know they try to make their money by being a top notch supplier of tiny motor control boards.
That could be used with your voltage level shifter as they often require pullup resistors to +3.3 V and +5.0 V.
The actual power for the board and the +5VDC come through whatever voltage is driving the motor.
++++++
In the past, I tried making my own H-bridges with BJTs and everything was very hot, required big heatsinks, and a bit larger than I wanted. This is tiny (but still runs quite hot) and seems to not require the heatsinks that take up so much space.
To me, that is a good value. Yes it runs two motors.
Please look on the photo of the board for Vccb > Vcca. Vccb would connect to +5 and Vcca would connect to +3.3.
So everything comes together in a tiny set up. And these boards are always used if you want to cannibalize the project and do something different. There appears to be lots of flexiblity in voltage and wattage.
You claim there are some useful BJTs for 10amp, 12VDC -- but don't specify anything. Maybe there are, but I don't have any idea what they are.
Try this one http://www.nxp.com/products/bipolar_transistors/low_vcesat_biss_transistors/low_voltage_low_vcesat_biss_transistors/low_vcesat_biss_transistors_single_npn/PHPT60415NY.html
That is sepc'd to 15A, and has 250mV drop at 10ADig into the details however, and the problem is not so much VceSAT, but where to find the 1A of base current.Extract that direct from 12V, and you have 12W of base-drive-power Even at 100% SMPS source at exactly the required 1.35V Vbe you have 1.35W in the base junction alone.Collector power loss at 10A is 2.5W
Given MOSFETS avoid the base-power issue entirely, and have lower Rds values, they are a more logical choice at 40V
You claim there are some useful BJTs for 10amp, 12VDC -- but don't specify anything. Maybe there are, but I don't have any idea what they are.
Try this one http://www.nxp.com/products/bipolar_transistors/low_vcesat_biss_transistors/low_voltage_low_vcesat_biss_transistors/low_vcesat_biss_transistors_single_npn/PHPT60415NY.html
That is sepc'd to 15A, and has 250mV drop at 10ADig into the details however, and the problem is not so much VceSAT, but where to find the 1A of base current.Extract that direct from 12V, and you have 12W of base-drive-power Even at 100% SMPS source at exactly the required 1.35V Vbe you have 1.35W in the base junction alone.Collector power loss at 10A is 2.5W
Given MOSFETS avoid the base-power issue entirely, and have lower Rds values, they are a more logical choice at 40V
"Dig into the details however, and the problem is not so much VceSAT, but where to find the 1A of base current."
Yes, I thought of that, but Eagletonium is getting fed enough confusion that I didn't want to get into the long drawn out debate on the feasiblity of DIYing a BJT H-Bridge. I spent a lot of time on the Parallax forums getting people to explain to me how to do and did a lot of reading elsewhere -- and then realized the result was a very hot, big power hog.
Peter J. must be mad at me as he has been throwing curve balls.