Desparate for some quick help

steprog

Hi Guys,
I am about to use this Freescale H-Bridge MC33886 for my design and I have less than an hour to send off the circuit board, but I decided to do a quick check and I think this device is 0-5volts instead of 3.3V.

If it is am I dead in the water with driving it with a prop?
Thanks,
Greg

T Chap

You could do several things. Maybe the device will switch with 3v3? Or, just drop a cmos 4050 inline and power the 4050 with 5v, the Prop pins will drive the 4050 and the output will be 5V.

steprog

I'm afraid I don't have much room left. I was thinking of putting pull-up to 5 volts, but with 12 lines this maybe too much as well. Who make devices that are only 0-5volts anymore? This isn't an old chip.

Microcontrolled

If you directly pull up to 5V, you will risk damaging the Propeller.

steprog

The Data sheet specifies a min of 3.5 Vdc for high. So I'm wondering.... is this good enough

Leon

No, you need 5V.

steprog

Hmm..
I guess wishful thinking on my part. It claims to be +5 volt cmos compatible afterall. Which should allow +3.3 to work. If it switches at the midpoint. However I think the ttl compatibility is messing things up.
Oh well.

steprog

Is the 4050 the best way to go?

jazzed

The pull-up is somewhat risky with 12 lines, and other factors including switching time constraints.

Your data sheet says Vih is 3.5+ volts. Use a 4.7K pull-up and let that produce the high level like open collector - the slew rate could be a problem, but your device input does have a 0.7V hysteresis. Propeller can drive low for the low level (12ma max total sink for 4.7K and 6ma for 10K pullup). If there are inputs to the propeller use a 2K series resistor - slower slew rate of course.

Standard disclaimer applies. It's your circuit and your decision.

steprog

No inputs just driving 3.3 to5volt

T Chap

Do you have room for a 3904 SMT transistor? How many signals?

steprog

I think this would do it as well. The protection diodes limit the voltage to 3.9volts I have been told. That should be okay. Not fast slewing here either. So 10k? any thoughts?

steprog

12 signals and not much space and not much time left

T Chap

steprog wrote: »

12 signals and not much space and not much time left

4050 is 6 gates each IC. 12 transistors will take slightly more space to lay out. Stick two 4050's end to end. If you don't have the space on the board, then add some headers and make an interface board with +5, GND, 12 signals to the smaller PCB, stick the small board in between. A 14 pin (2x7) header on .1" centers will allow the 14 connections, then use a 2x7 ribbon header or similar to connect the board.

If you can't manage the space or the extra board in the order, then just add some pads for +5, GND, 12 connections. Go to radio shack and get the little hobby perf boards and make up the 4050 board with through hole parts, very simple solution for a one-off. If you use .1" center pads for the 12 pads, then you could even solder the smaller board onto the header pins.

steprog

I guess no one likes Jazzed solution. I would think that if you limited the current on the protection diodes you would be good. no?

jazzed

I would be more concerned about input problems than output problems. The protection diodes are only on the inputs AFAIK.

I use the pull-up trick on I2C with propeller pins with no problems. I do have a series resistor between devices even on the Propeller outputs though - it doesn't hurt and can save your butttt.

Silence doesn't necessarily mean disapproval around here. This will be certain though: if your board doesn't work for some reason either immediately or over time because of using my suggestion, I'll probably take s**t for it though even with the disclaimer. Like I said, your design.

steprog

I understand. It's amazing how much stuff you can catch.

steprog

If I remember right there is a pullup resistor for the eeprom right?

hinv

What about driving your propeller at the upper limit of 3.6V?
I realize driving on the edge is probably bad for production applications, but it may get you through a pinch. ...and that propeller is pretty resiliant.

steprog

not a bad thought.

jazzed

steprog wrote: »

If I remember right there is a pullup resistor for the eeprom right?

Yes, but it's normally pulled to 3.3V. The PS2 connectors on the Propeller Demo Board use pullups to 5V via 10K, but they are normally used for Propeller inputs. Seems to work fine for the brief times the pins are outputs.

Beau Schwabe

Lift the ground supply to ONLY the Propeller by 1 diode drop!

That way when the I/O goes 'HIGH' , instead of 3.3V with reference to the 'actual ground' everything else sees, you should 'see' about 3.9V ... When the I/O goes 'LOW' you would 'see' the diode drop ... about 0.6V

steprog

This looks interesting. Can this screw me up with other systems hooked up to me. I assume all grounds would need to be lifted. If I have a +3.3v adc connected will that throw that off? My guess is maybe.

Beau Schwabe

steprog,

In that case with the ADC or anything else requiring 3.3V use the Version#2 of what I posted. As far as the Propeller is concerned (and anything else connected to the regulator in this way) it will only 'see' 3.3V directly from the regulator. You still need to use the normal caps on the regulator.

Not all grounds... just the ground to the Propeller and any other 3.3V in communication with the Propeller.

steprog

I like this, but this might have unintended consequences. What happens to all the inputs from other devices that are 3.3 volts. I like it I just trying to see how it would affect the whole picture.

steprog

Do you think this is better than just putting some 10k pullups? I have plenty on this board for instance I have instrumentation amps(+5Vdc) hooked up to an adc(+3.3V) that is connected to the prop.

Beau Schwabe

lol - you ask a layout guy if it's ok to forward bias a diode meant for ESD protection on the I/O ? - funny :smilewinkgrin:

With a 10K Pullup resistor your pulling the I/O to about 3.88V and forward biasing the HIGH side ESD diode built-in to the I/O Pin. It's only 112 uA, but it's "ON" and you increase the potential for CMOS latchup. (See attached PULLUP.JPG)

If my original solution won't work, It was more of a 'fix' after the fact, my personal preference is to use an Emitter follower. (See attached Emitter Follower v2.JPG). With the Emitter Follower, you avoid forward biasing the ESD diode, and you get a higher voltage of 5V from the pull-up resistor.

steprog

Thanks Beau, I don't mean to challenge you, but sometimes to get a better explanation it helps. Besides who better to ask...
Thanks for your help on this.
Greg

steprog

I finally decided to put 2 cd4504 level translators on this board and delayed my board a day. I got it in to do the level translation, but man its tough to find info on that device and how to use it. Anyone used this device?

steprog

The more I look at this device the less I like it. No documentation not real clear anyway and it doesn't seem to be spec for 3.3 to 5 volts. It still might work but does anyone have a better choice?
Thanks,
Greg

steprog

T_Chap,
I think I finally cycled back to using your cd4050 ideal. There maybe better translators out there buy at least I understand this.
Thanks,
Greg