Post by thread:[EE]: Darlington Speed-trap

I was testing the ULN2803 mentioned in another thread, and was amazed at how fast this chip is. I drove a 130 mA load 30 ohm resistor [5v - Vcesat(=1v)] at 2 mhz. Nice and clean. For those not familiar, the 2803 chip is an 8-channel darlington driver, rated at nominal 0.25usec turn-on/turn-off delays [ala 50-50 points] --> and it actually does this under load. Anyone else have experience using these chips at high-frequencies under load? I thought darlingtons were supposed to be slow. thanks, - dan michaels http://www.oricomtech.com ========================== -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Dan: Yes, I have seen similar times. I once tried using them to pulse a 5MHz (yes, MegaHertz) transducer with the '2803. Seemed to work. Never pursued it doe to other factors. As you may know, Sprauge (Allegro) spec's the turn-on time at 250ns typ and 1us max. They "forgot" to give a test circuit. They don't spec a rise or fall time at all. One observation I have is that the smaller the load resistance, the better the rise and fall times should be. This is because of lower gain and therefor less Miller effect capacitance. Another feature of these parts is that under low loading, they have low saturation voltages. This means that they can be used as "protected" buffers to "fragile" PIC input pins. -- Rich Dan Michaels wrote: {Quote hidden}> > I was testing the ULN2803 mentioned in another thread, and > was amazed at how fast this chip is. I drove a 130 mA load > 30 ohm resistor [5v - Vcesat(=1v)] at 2 mhz. Nice and clean. > > For those not familiar, the 2803 chip is an 8-channel darlington > driver, rated at nominal 0.25usec turn-on/turn-off delays [ala > 50-50 points] --> and it actually does this under load. > > Anyone else have experience using these chips at high-frequencies > under load? I thought darlingtons were supposed to be slow. > > thanks, > - dan michaels > http://www.oricomtech.com > ========================== > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Several Allegro chips probably drive half the bubble printer heads in the civilized world right now. That's pulses down to 1usec. Several older switcher designs used Darlingtons. Darlingtons are not slow when designed not to be. In theory the Ton/Toff of a Darlington is the sum of the TonToff of the component transistors. In practice it is slower because of capacitance (esp. Toff). This is why most Darlingtons have a built-in resistor network near the base of the second transistor (including in the ULN2803). The worst problem with Darlingtons is Vcesat. Try to run a ULN2803 at max spec and you will exceed SOA. I have a small stepper driver (experimental) where I had to glue a small fin on the ULN2803 to cool it while driving a motor with Iin (total) 450ma at 12V. With three outputs on at max spec you exceed SOA (power dissipation). As long as you keep this in mind, you will be fine. Peter -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spam_OUTlistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

> I was testing the ULN2803 mentioned in another thread, and > was amazed at how fast this chip is. I drove a 130 mA load > 30 ohm resistor [5v - Vcesat(=1v)] at 2 mhz. Nice and clean. > > For those not familiar, the 2803 chip is an 8-channel darlington > driver, rated at nominal 0.25usec turn-on/turn-off delays [ala > 50-50 points] --> and it actually does this under load. I am using a ULN2004 (7 channel relative to ULN280x) with pulse widths down to around 1 microsecond using the active pullup mentioned below. Word of warning for those reading this and thinking they can use it as a "fast" driver with a resistive load (my active pullup cures this). The ULN2803 is of course an "open collector" driver. There is an active pull down (darlington pair) but NO pullup whatsoever. The turn off time will be no faster than the turnoff of the driver (which Dan is commenting on) BUT the actual time for the output voltage to rise will then depend entirely on the characteristics of the driven circuit. If you use a highish value load resistor and drive a capacitive load you MAY get a time constant of far longer than the driver turn off time. This is just basic electronics but is an easy trap to fall into if you are used to using a standard push-pull output in which there is a low and a high output driver thus providing a low impedance on and off drive circuit. A ULNxxxx driver (or any open collector driver) can be converted to an active on and off driver but the addition of a transistor, a diode and one resistor. While this reduces the advantage of the single chop drive there are circumstances where this is a very useful approach. In one design I use this on one channel of a 7 channel driver to drive a power FET at high speed. Parts: NPN transistor for hi side driver (eg BC337 TO92 package) Diode (say 1N4003) Resistor (1k to 10k range) Transistor collector to positive supply Transistor base to ULN2803 output Transistor emitter to output. Diode cathode to base and anode to emitter (ie diode conducts to pull output LOW when ULN28-3 output goes LOW). Resistor from transistor base to positive supply. While this circuit features inside many ICs I was first introduced to it as a cheap and simple FET driver using discrete transistors by AFAIR Lance Allen. I have never seen it used with a ULNxxxx input but it is an obvious extension of the basic principle. Turnoff time is now mainly controlled by the ULNxxxx to transistor drive time constants and can be designed to be as fats as required essentially regardless of load. It's main shortcoming (especially here) is that the off/low output voltage is 1 diode drop plus ULNxxxx saturation voltage above ground. This would be a drawback in some circuits but is unimportant when driving MOSFETS. Apart from power wasting and loss of available voltage this is not usually a problem when driving eg motors, relays etc. Russell McMahon -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body