A better H bridge...

[Hero999]

I can easily find TIP transistors, and they are much cheaper.

How much cheaper?

Where I live cheap MOSFETs are the same price as BJTs.

 

[Roff]

Thanks everybody,

now here is the final version, after all your comments:

**broken link removed**

Leave the logic part.. i know its fine, please just concentrate on the red shaded part and veryfy its free of errors.

I calculated all the resistances so that the base current in the 4 TIP transistors be around 38 to 43 mA.

thx a lot

Are you sure about the logic? It looks to me like you designed it for NPN pullups (Q5 and Q7). Keep in mind that NPN pullups don't invert, but PNPs do.
I think you now have a recipe for smoke.

 

[i_build_stuff]

Here are the logic functions I came up with, for the base drives of Q1 - Q4:

b_Q1 <= P1 and not P2 and PWM;
b_Q2 <= not ( P2 and PWM );
b_Q3 <= P2 and not P1 and PWM;
b_Q4 <= not ( P1 and PWM );

I'm not sure if this is exactly the same as the original, but the inputs for different actions would be:

P1P2 -- action

00 -- freewheeling
01 -- forward
10 -- reverse
11 -- braking

(disclaimer: I'm half asleep right now, so you might want to run it through a simulator first)

 

[Roff]

I haven't used PICs, but I think that controlling U1 through U4 independently from 4 ports on a PIC would be more versatile. You will need to build in dead times to allow for transistor storage times (turn-off delay), which would otherwise cause high shoot-through currents which can potentially destroy your H-bridge transistors.

 

[ikalogic]

Here are the logic functions I came up with, for the base drives of Q1 - Q4:

b_Q1 <= P1 and not P2 and PWM;
b_Q2 <= not ( P2 and PWM );
b_Q3 <= P2 and not P1 and PWM;
b_Q4 <= not ( P1 and PWM );

I'm not sure if this is exactly the same as the original, but the inputs for different actions would be:

P1P2 -- action

00 -- freewheeling
01 -- forward
10 -- reverse
11 -- braking

(disclaimer: I'm half asleep right now, so you might want to run it through a simulator first)

Thanks you that's exactly what it does..

And PLEASE! guys i told you the logic is right.. i tested it on many other bridges.. and Ron H, there wasen't any smoke at all.

Now PLEASEEE just concentrate on the red shaded part PLEASE! i know the logic is fine, and as 'I_build_stuff' said, there are 4 states :

00 -- freewheeling
01 -- forward
10 -- reverse
11 -- braking

Now all i am asking, to all professional guys here is very simple:

again, I calculated all the resistances so that the base current in the 4 TIP transistors be around 38 to 43 mA. are there any ANALOG related remarqua about the transistor saturation...

i just need a confirmation on this part only!

i SWARE on everything holly i beleive in that the logic part is fine and was tested in many project at my university and worked very fine.. i am just trying to increase the performance of the device..

again thanks.

 

[Roff]

Consider this:

 

[Roff]

Regarding your bridge-related questions - we need to know the load current before they can really be answered.

 

[ikalogic]

Regarding your bridge-related questions - we need to know the load current before they can really be answered.

You were right, absolutely, but this was a simple mistake i made when implementing the PNP transisstors.. i forget to change the way the 2 other NPN trtansistors are driven..

So here i corrected the circuit

**broken link removed**

So what do you think now Ron H? thanks a lot for your help.

The expected load current wont exceed 3 A.

Anyother bugs detected? , should i go on producing this H bridge?

 

[Roff]

It's still going to smoke. More on the resistor values later.

Edit: I forgot to take the 220 out of the Q4 collector circuit. This doesn't change the fact that it won't work.

 

[ikalogic]

still red faced... but i guess that's how we learn..

It's still going to smoke. More on the resistor values later.

Edit: I forgot to take the 220 out of the Q4 collector circuit. This doesn't change the fact that it won't work.

Again Thank you, you'r saving me big troubles.

Now here is a different approach that should solve it, PLUS, as u seemed more enthousiastic about 'Active Turn-Off' in this design all the TIPs are driven by active turn off.

**broken link removed**

what do you think now? i guess there is no more smoke.

About the resistor.. well i will use 1/5 watt rated resistors.. and after some calculations, power dissipated though resistors shouldn't ever exceed 0.3 watt.. it that dangerous?

EDIT: okay, the 220 resisitors, may disipate more than 1/5 watt.. about 0.65 wat.. are 1 WATT rated resistors going to solve the problem?

 

[Styx]

why is Q8 being driven differently?

 

[Roff]

Your emitter follower outputs (Q1, Q3) will only go to about 3V. They need to go to 12V, which means the base will need to go to about 12.7V.
Your resistor values are lower than they need to be. You only need base drive of ~Ic/250,or about 12mA.
You need dead time during switching, as I mentioned in a previous post.

 

[ikalogic]

why is Q8 being driven differently?

was a mistake, that was corrected right now,

i am working on on ron h advices, and will be uploading another schematic soon.

 

[ikalogic]

Your emitter follower outputs (Q1, Q3) will only go to about 3V. They need to go to 12V, which means the base will need to go to about 12.7V.
Your resistor values are lower than they need to be. You only need base drive of ~Ic/250,or about 12mA.
You need dead time during switching, as I mentioned in a previous post.

Yes, i understand this point.. i don't know how you manage to detect errors that fast in a circuit.. i guess its a matter of experience.. anyway,

now here is the only solution i see to that specific problem, but sadly this would mean to lose the Active-turn off capability..

For the deat time, its noted. but when you say dead time, you mean a time where all transistors are OFF before running the bridge in the oposit current direction? to make sure no short circuit happens right?

And for the resistor.. the datasheet says i need 30 mA.. sometimes more.. anyway, i'll relie on testing to adjust the exact values..

But most important, did this solve the problem for (Q1, Q3)?


thx for your patience ron h

**broken link removed**

 

[ikalogic]

Ron h is gone, i think he's saturated by my questions today.. anybody else ready to help where he left?

 

[Roff]

Yes, i understand this point.. i don't know how you manage to detect errors that fast in a circuit.. i guess its a matter of experience.. anyway,

now here is the only solution i see to that specific problem, but sadly this would mean to lose the Active-turn off capability..

For the deat time, its noted. but when you say dead time, you mean a time where all transistors are OFF before running the bridge in the oposit current direction? to make sure no short circuit happens right?

Yes, because the transistors do not turn of instantaneously. I ran a sim that suggested you should turn one half of a leg off about a microsecond before you turn the other half on.

And for the resistor.. the datasheet says i need 30 mA.. sometimes more.. anyway, i'll relie on testing to adjust the exact values..

Where did you find this? The OnSemi datasheet has all saturation curves with IC/IB = 250.

But most important, did this solve the problem for (Q1, Q3)?


thx for your patience ron h

No, the p-p output swing of an emitter follower (Q1, Q3) will always be less than the input swing.

 

[ikalogic]

Yes, because the transistors do not turn of instantaneously. I ran a sim that suggested you should turn one half of a leg off about a microsecond before you turn the other half on.

Noted. thx

Where did you find this? The OnSemi datasheet has all saturation curves with IC/IB = 250.

Also noted, i'll make the base current about 20 mA max...

No, the p-p output swing of an emitter follower (Q1, Q3) will always be less than the input swing.

here a litle change that may solve this problem.. if i didn't get it right this time.. may you please give me some instructions to overcome/decrease this problem..

**broken link removed**

 

[ikalogic]

forget it ron h, this circuit wont work either... don't waste your time on this one.. i'll have to do a new design from scratch.. bearing in mind all what i learnt here.

i guess i'll have to use a different logic gates arrangement.. anyway.. we'll see.. see you in a day or 2 with a new design.

thx for everything.

 

[ikalogic]

just one last hint for today:

http://www.mayothi.com/transistors.html

please tell me this guy is wrong about PNP transistors, when he says that connecting the base of a PNP to ground means no current is flowing !!? shouldn't there be current flowing out from the base?? right?

 

[Nigel Goodwin]

just one last hint for today:

http://www.mayothi.com/transistors.html

please tell me this guy is wrong about PNP transistors, when he says that connecting the base of a PNP to ground means no current is flowing !!? shouldn't there be current flowing out from the base?? right?

Yes, he's talking complete rubbish - there's no difference between NPN and PNP, except their polarity.