Shop OBEX P1 Docs P2 Docs Learn Events
Building a proper h-bridge from scratch - Page 2 — Parallax Forums

Building a proper h-bridge from scratch

2»

Comments

  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-08-27 09:58
    [QUOTE=Circuitsoft;1121638]Not really. The lowest dissipation will happen when the transistor is held just on the edge of saturation. Saturation happens with the Base-Collector diode becomes forward biased and draws current off the base. The B-C diode always has a lower forward voltage then the B-E diode, which is why saturation can even happen. So, the saturation voltage of C-E on a bipolar transistor is actually often around .1-.2 volts. It's not a diode drop. Anyway, if you're pushing a beta of 10 or less, the power coming into the transistor from the base actually becomes significant, and you can dissipate a lot of heat in the base-collector junction, which is not made to take much forward current. It's especially important to keep transistors out of saturation when doing PWM because the switching time is much faster if it's not saturated. That's why I assumed a beta of 15-20 in my circuit; that number (20) came straight out of the BC140 datasheet for 1A collector current. By feeding it enough current to just make the B-C junction start to forward conduct, you get much better switching speed because fewer charge carriers need to be neutralized for the current flow to turn off.[/QUOTE]

    I see that the TIP120/TIP125 and the TIP35c/TIP36c perform much better in regards to saturation voltage drops at less that their highest rated continous output.

    The TIP120/TIP125 really should target 3 amps, not 5amps output as the voltage drop is just 2 volts at 3 amps compared to 4 voltage at 5amps. This is about using some number crunching AND learning to read all the details in the PDF.

    The TIP35C/TIP36C really should target 15 amps, not 25 amps output as the voltage drop is just 1.8 at 15 amps compared to 4 voltages at 25 amps.

    And, the current limiting resistors for the Base current should likely be fitting tightly to an actual motor to optimize the low end of the saturation current.

    I suspect that Bob Blick's design was successful because it does handle larger than average motors quite well, but really trying to use it to the max requires someone to spend quite a few bucks for an 1/8 hp motor and that doesn't happen all that often.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-08-27 13:00
    Mark_T wrote: »
    True but not very usable - the edge of saturation varies with device and temperature and you'd have to dynamically measure the saturation voltage to be able to
    maintain this state - the small excess dissipation involved in being more saturated than this point for all devices/temperatures is a lot smaller than the excess
    dissipation should a device go below saturation (even when driven hard the collector circuit is taking 10 times the current of the base so that 0.1V at the collector is
    worth 1.0V at the base).
    Dissipation is more of an issue than you'd think. There is no diode drop from C-E, so the only reason it doesn't go to 0 volts is because the B-C diode forward biases and starts drawing current off the base. So, the C-E voltage on a saturated transistor can be as low as .1v, and B-E can be .7 to 1v. So, even with 1/10 the current going into the base, that makes almost half the power dissipation.

    As for holding it on the edge of saturation, that's actually pretty easy. Since hFE drops like a rock when you hit saturation, you adjust your base current for about half of that drop (I plan on hFE 15-20) and the transistor will regulate itself to the edge of saturation. If the load is more dynamic, you can use a multi transistor setup, or a dual-emitter transistor.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-08-28 01:30
    There really seem to be two separate design situations - with PWM and without PWM. Obviously, one can operate with a higher motor drive voltage and with less limitations to saturation if one uses PWM. In fact, one may have to.

    But, I was simply trying to design a reversible motor use that would not use PWM - let's say using a motor to open and close a window for environmental control. One needs some power, but not the refinements of PWM.

    Still, the issues of how and where heat build up inside the transistor junctions are quite informative. And PWM can better manage the issues of heat.

    I know transistors do vary quite widely in targeted qualities. Can these be tested and the h-bridge designed to particular stats?
  • jmgjmg Posts: 15,173
    edited 2012-08-28 14:32
    I know transistors do vary quite widely in targeted qualities. Can these be tested and the h-bridge designed to particular stats?

    Individual testing ? Anything is possible, but that is very expensive to start with, and costly to service.
    I know a company that tests Lazers, at great cost, but cannot imagine anyone would bother testing transistors, when another few cents can easily buy a better part.

    As a reality check, I found these parts for Simple Motor Drive (as in your window example) :

    A4953ELJTR-T 620-1428-1-ND 1 $1.95 50 $1.57 100 $1.41 Stock: 6,174
    Description IC PWM MOTOR DVR FULL 8SOIC 2A 40V

    A4950ELJTR-T 620-1400-1-ND 100 $2.25 Stock: 8,809 Pin compatible.
    Description IC PWM MOTOR DVR FULL BRDG 8SOIC 3.5A 40V

    These are H bridge, SO8, and they have a current-sense internal PWM mode.
    For Motors inside their V=I operating area, they look hard to beat.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-08-30 23:36
    Can one just simply take 4 solid-state relays and configure these into a working h-bridge?

    You would have opto-isolation, and the ability to handle both high current and high dc voltages. No much to build.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-08-31 08:21
    SSRs tend to have pretty high voltage drop, but yes.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-08-31 08:43
    Hi Loopy;
    @everyone
    Fly back diodes are assumed. After a bit of reading between conventional rectifier diodes, geranium diodes, and Schottky diodes - it seem Schottky diodes are the most appropriate - something like a 1N5818.
    Schottky and germanium diodes are generally not appropriate in fly back applications. They have lower than normal voltage drops. This only lengthens the time to dissipate the magnetic energy in the coil.

    What you want is a diode with MORE voltage drop. Even to the extent that one uses a zener diode for more voltage. The effect is to dissipate the the coil current more rapidly.

    Some may add a carefully chosen resister in series with the fly back diode.

    The intention here is to raise the dissipation voltage but not so high a voltage as to exceed the rating of the driver transistor.

    Duane J
  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2012-08-31 16:10
    I used to make very ugly, freeform bridges using four pn2222 and two pn2907 transistors.
    They were just to control tiny motors so the limited current rating didn't matter.

    Only trouble with the simple design was if you somehow managed to have both driver
    pins on the uC high at the same time then smoke would quickly come :-)
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-09-01 00:26
    Hi Loopy;

    Schottky and germanium diodes are generally not appropriate in fly back applications. They have lower than normal voltage drops. This only lengthens the time to dissipate the magnetic energy in the coil. I'll use the Schottky for blocking diodes, where I need less voltage drop.

    What you want is a diode with more voltage drop. Even to the extent that one uses a zener diode for more voltage. The effect is to dissipate the the coil current more rapidly.

    Some may add a carefully chosen resister in series with the fly back diode.

    The intention here is to raise the dissipation voltage but not so high a voltage as to exceed the rating of the driver transistor.

    Duane J

    Thanks, I guess engineering always trumps the sales pitch in the PDF. I thought the lower forward voltage drop would be a plus.

    @Holly
    I have been using some 2n2222/2n2907 freeform h-bridges based on the Tilden bridge design. At first I thought these were really great, but they got very hot in use. They work on a small scale and the linear region may be ideal for BEAM.

    Now I have a couple of L298 h-bridges that are being used instead. These are good, though a bit expensive in kit form. So I was trying to improve my mathematical understanding of what to use and how.
  • ben_r_ben_r_ Posts: 11
    edited 2013-03-15 11:28
    The ZXMHC3F381N8 is a nice H bridge, good to 2.5 amps (3 amps for a 10s surge) with gates driven at 5V, and 3 amps (4 amps for 10 seconds) with gates driven at 10V. A pair of FAN3268 are a little on the expensive side, but will solidly drive this H bridge at 5 or 12v (up to 18v).

    For smaller motors, you can drive that same H bridge with just a bus driver such as the 74LVX4245.
    Soo glad you posted this! These types of low cost, SMD, single H-bridge driver ICs were exactly what I was looking for! Thank you!
  • mibarra3mibarra3 Posts: 4
    edited 2013-11-03 19:20
    hi, I am new in this forum. I am building an h bridge circuit to control a 12 v 7 amps dc motor and need help to calculate the resistors and diodes. can you help me?
  • ben_r_ben_r_ Posts: 11
    edited 2013-11-03 19:22
    Whoa! Youre trying to drive a 7A motor?!
  • mibarra3mibarra3 Posts: 4
    edited 2013-11-03 19:31
    hi, I am new in this forum. I am building an h bridge circuit to control a 12 v 7 amps dc motor and need help to calculate the resistors and diodes. can you help me?
  • mibarra3mibarra3 Posts: 4
    edited 2013-11-03 19:35
    thanks for answering so fast. yea I got some TIP141. but not sure what resistors and diods I need, can you help?
  • ben_r_ben_r_ Posts: 11
    edited 2013-11-03 21:43
    mibarra3 wrote: »
    thanks for answering so fast. yea I got some TIP141. but not sure what resistors and diods I need, can you help?
    Unfortunately I have never switched loads that big, so off the top of my head I dont think I can help.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-11-04 03:33
    Hi,
    Well this is a favorite topic. You can't build a 7 amp H-bridge with Darlingtons, but you also need both NPN and PNP darlingtons. The TIP141 are just the NPN. (See futurelec.com for a list of available TIPxxx)

    Reality is that the TIP141 is only rated to 5 amps, 7 amps is too much. And even if you could use TIPs, they are going to run sizzling hot in comparison to power MOSfets.

    If you can double the voltage to the motor and halve the amps, those Darlingtons might work as they have a higher voltage ablilty. Instead of a 12 volt motor at 7 amps, run a 24 volt motor at 3.5 amps.

    SO, if you are going to have to start over from scratch... it would be best to look at a Power MOSfet solution that to bother with TIP141s.

    For Bipolar at 7 amps, you would need the TIP3055 and TIP2955 which can handle up to 15 amps as your final stage. But, they are NOT Darlington's so you would have more design details to work out.

    For a 7 amp motor, you really should have about 15amps rated devices. Why? Inductive loads and stalled motors are going to spike way over the 7 amp nominal motor load.

    So think about all this.
    If you really want to build a TIP141/TIP145,6,or7 H-bridge, you can copy Bob Blick's design that is intended for TIP130/TIP135 with the same resistor values and his schematic. It should work fine. BUT the 5 amp rating really means you should run a motor that at the very least demands 1 amp less.

    Driving bipolar transistors to their limit will cause a thermal runaway that destroys the device. Due to thermal runaway problems, never put BJT into a parallel configuration to get more power.

    Mosfets do NOT have a thermal runaway. And you can actually use several MOSfet in parallel to get higher power
  • Mark_TMark_T Posts: 1,981
    edited 2013-11-05 09:08
    I'd never contemplate the complexity of driving bipolar transistors for a big H-bridge, its not trivial.

    MOSFETs or IGBTs every time. My favorite MOSFET driver at the moment is FAN7388 or FAN7888 which is
    3-phase (so it can drive an H-bridge with a channel spare), and its a few dollars. But there are others and many
    are rated at 200V or even 600V...

    The main thing you have to realize with high power H-bridges is that you spend much of your efforts on designing
    the protection circuitry - undervoltage shutdown, protection from transients, and gross-overcurrent shutdown need to be
    there. Undervoltage shutdown is standard on MOSFET driver chips though, and most have either fixed or tunable
    dead-time to prevent shoot-through.

    When higher voltages are involved IGBTs start to make sense (much more rugged than MOSFETs since the drain voltage
    is kept away from the delicate gate), although they have a large Vsat loss (2+V) that is only acceptable when the supply is 100V
    or more. It never makes sense to attempt MOSFET circuits above that voltage these days, the gate-drain capacitance is a real
    weakness with that sort of output voltage swing and the gate oxide is the weak point in any MOSFET.

    My biggest H-bridge to date is using 100V 200A devices, bolted to a block of aluminium, and expected to handle upto 50V at 50A
    it should be coasting.... Big 15V protection zeners are mounted right on the gate/source terminals as the first level of protection, then the
    drivers are powered from a DC-DC converter to allow them to float w.r.t battery ground (I expect serious ground offsets at those
    currents) and to handle battery voltage dips.
    Opto-isolated inputs to the drivers protect the control logic should the thing fail, and reduce susceptibility to noise, and allow
    control logic to use a different ground.
    The battery and motor wiring is 10 mm^2 and even then shows as much resistance as the devices themselves, perhaps upto 2V
    lost across MOSFETs and wires out of 24V (current testing setup). Whole thing is destined to power a golf-buggy transaxle in
    the 1hp range.

    So my other advice is over-engineer!
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-11-05 10:14
    Over-engineering is quite necessary with induction fields in motors, a lot of junk created by motor brushes, and the potential for near stall startups under heavy loads.

    I was very stubborn about wanting to build a BJT H-bridge, but the realities of how far power MosFets have progressed make it just an exercise of going down memory lane. It is even getting near impossible to buy high amperage BJT devices as the market that was once driven by transistors to control fly-back transformers in CRTs is nearly gone.

    There is a thin market in audiophiles that desire to build Class A transistor amplifiers, but trying to get the power transformers and the filtering capacitors to complete such a project is drying up as well.

    Everything is going or had gone mosfet and digital. Recently, I tried to buy a tuning fork to tune my guitar and all they had was digital tuning devices.. the salesperson couldn't grasp what a tuning fork was or might even look like. He acted like I was asking how to shoe a horse.
  • mibarra3mibarra3 Posts: 4
    edited 2013-11-05 20:07
    thanks all of you for your very appreciated comments and suggestions, I think I now have a better understanding of what am doing. one more question, where can I find the Bob Blick's design intended for TIP130/TIP135 that you mention in your comments.?:innocent: thanks and "talk" to you all latter..
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-11-06 07:40
    mibarra3 wrote: »
    thanks all of you for your very appreciated comments and suggestions, I think I now have a better understanding of what am doing. one more question, where can I find the Bob Blick's design intended for TIP130/TIP135 that you mention in your comments.?:innocent: thanks and "talk" to you all latter..
    H-Bridge
    This uses the 5A TIP120/TIP125 Darlingtons.
    Just substitute the 8A TIP130/TIP135 Darlingtons. Should work fine.
    However, MOSFET designs work much better.

    Duane J
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-11-06 11:03
    Hmmm.. I feel so stupid. The TIP120/125 series are 5amps, the TIP140/145 series are 5 amps at a higher voltage, and the TIP 150/155 are 5 amps at an even higher voltage.

    But I forgot to check the TIP130/135. Part numbers can be so confusing.
Sign In or Register to comment.