Hi kal,
Once again, being confused, as you put it, is a good thing (not meant to be condescending). The reason: R35 (5K6) and R24 (470R) do not do anything in terms of the basic function of the circuit at DC and are really a nuisance, especially R24. So, why are they there?
(1) R24 is only there for protection:
(1.1) if the current transfer ratio (CTR) of the particular 4N25 fitted is very high, R24 limits IE from the opto receiving transistor (OTR) to around 24V/470 mA. That current is on the high side, but I did not want to make R24 to big as it slows Q27 turn off.
(1.2) Also R24 limits the current, as mentioned for previous circuits, in the case of extraordinary events.
(2) R35 serves two purposes:
(2.1) R35 makes sure that any leakage current (dark current) from the ORT does not turn Q27 on
(2.2) R35 provides a discharge path to 0V for the voltages stored on the parasitic capacitances, both real and virtual, at the base of Q27. Without R35, Q27 would turn on faster, but it would take an age to turn off. The ORT would also take an age to turn off. In fact, in the interests of speed it would be good to make R35 as low as possible: 100R would be ideal, but 5K6 is as low as I could go and still have the ORT drive Q27 sufficiently, in the case of a low CTR 4N25.
(3) Speed
As you intend to switch the motor on and off at relatively high frequencies, speed become an important aspect of the design. Specifically, in this application, how fast the transistors and optocoupler will turn on and off.
(3.1) Optocoupler
While on the subject of speed, the 4N25 optocoupler, although an industry standard, is fairly slow. Had I been fully aware of the speed requirements I would have probably specified the much faster HCPL4562 optocoupler, even though it is more expensive (£2.50 UK). Incidentally, operating an optocoupler ORT in common collector configuration, as in the present circuits, is faster than common emitter.
HCPL4526 Data Sheet: https://www.avagotech.com/docs/AV02-1361EN
(3.2) Logic Gate
The latest circuit of post #40, with the logic gate in the micro side, is much, much faster and better characterized than saturating transistors and is thus the recommended approach. Arguably, it is also the simpler approach.
(3.3) Capacitance
In a nutshell, capacitance generally slows a circuit because capacitors have to be charged and discharged with a current in order to generate a voltage. The lower the capacitance and the higher the available charging/discharging current, the faster the circuit. Incidentally, voltage saturating a transistor collector also slows it. The more excess base current, the longer a BJT takes to turn off.
Even though all the transistors, including MOSFETs, in your circuit are theoretically either on or off, in the finite time when they are transitioning between the two states they are acting as common emitter/source linear amplifiers. While in this transition region, the MOSFETs will be dissipating significant power, so it is important, for efficiency, that the MOSFEts are snapped on and off as rapidly as possible.
What is a parasitic capacitance? There are unintended capacitors, inductors and resistors literally everywhere in the real world. The theoretical schematics that you see, in no way represent the actual physical circuit. This is why simulators can be very misleading if not used carefully. All this may sound complex and, when you get into very high frequencies, it is. But, at frequencies below about 10K Hz (it varies), you can generally ignore parasitics. With a BJT the two most important internal parasitic capacitors are Cbe and Ccb. If you look into the base of a BJT, the input capacitance, is very roughly, Cbe + (A*Ccb), where A is the voltage gain of the BJT in common emitter configuration. How do you know what A is. A BJT has an internal emitter resistor re which roughly = 25/Ic in Ohms, where IC is in mA.
So suppose you had a BC546 configured as a common emitter amplifier with a collector load of 2K2 + 2K2 = 4K4 Ohms and a collector current of 2.73mA, re would be 25/2.73 = 9.16 Ohms. The voltage gain of a common emitter BJT amplifier is RC/re. Thus, the voltage gain would be 4K4/9.16 Ohms = 480 (it would not be quite that high in practice but do not worry about that for the moment). This means that Ccb would appear (virtually) at the base as Ccb * 480. The Ccb of a BC546 would probably be around 10pF. so the virtual base capacitance would be 480 * 10pF= 4.8nF: massive in fact. This multiplying of the the Ccb is known as the Miller effect, after the chap who defined it way back in the valve (tube) days.
In common cathode/emitter/source amplifiers, the Miller effect dominates the input capacitance of the amplifier. Power MOSFETS, with their massive Cdg are quite difficult to use as amplifiers at higher frequencies. For example, a power MOSFET could have a Cdg as high as 500pF.
(4) General Note
As a general note, just be aware of speed issues, but don't let them distract you from your initial objective of fully understanding the basic DC circuit function. Just in case you feel overwhelmed by it all, as I did, be assured that it is not all that complicated, at a working level that is.
Last edited:
