Maker Pro
Maker Pro

My life was a lie

Daniel_B

Apr 30, 2021
7
Joined
Apr 30, 2021
Messages
7
Hi,
I just want to tell you about my story with transistors, because i really do feel stupid now.

I've been using various transistors for a long time. I always put them most of the time before the load because it seemed logical and that's how i've seen it elsewhere. I also supplied the gate/base with a voltage equal or smaller than the supply voltage. This became a habit and although i knew about the difference between potentials and voltage, the use of transistors this way seemed so straightforward that i never thought about it more deeply. So, until now i believed those 0.7 V or in case of MOSFETs 1-3 V losses were perfectly normal and couldn't be dealt with. I thought about it as a fact, something that can only be solved with a relay.
Though i had doubts, I really didn't think the circuit design was wrong. Maybe I'm just not using the correct MOSFET suited for the task, that's why they have slightly bigger losses (and i still thought that even better ones should have a voltage drop of at least 0.3-0.7 V). I've been using my transistors this way for at least 5 years with this belief. I used many simulators and i built many real circuits this way. It was not really a problem in signal processing and in some cases transistors do need to be used this way, but it caused massive losses in power applications.

I only noticed today that something is wrong. I started designing a PWM circuit that drives a motor. Because the electricity comes from the 230 V AC network, i had to do a lot of things with it first and as a result i had a quite few different potentials, which served as a reminder that potentials really are relative. As usual, i've been struggling with the voltage drop in my simulation program and this time it was really big, so i was playing around with the circuit. Eventually I increased the voltage at the gate/base of the transistor, which decreased the voltage drop. At first, I didn't understand why this is happening since i already supplied it with a quite high voltage. It is only then that the gears in my head started to turn. I had positive voltages on all sides of the transistor when it was on, so i had to think about it again, what voltage is really responsible for switching the transistor. Then I finally realized the error.
Thinking now about it, this behaivour is really self-explanatory, I don't even know how i didn't notice this until now that my transistors were only partially switched on. This is kind of embarrassing but it is also a pleasant suprise. I can now make improvements to power efficiency.
 

Audioguru

Sep 24, 2016
3,656
Joined
Sep 24, 2016
Messages
3,656
I guess you never looked at the datasheet of any transistor that says for saturated switch, the base current must be 1/10th of the collector current even when the hFE is very high. You mentioned a 0.7V loss then maybe your load was at the emitter instead of at the collector.

I guess you never looked at the datasheet for a Mosfet that says for most Mosfets, it conducts only 0.25mA (almost turned off) when the Vgs is 2V to 4V and it is fully turned on when its Vgs is 10V.
The on-resistance of an ordinary IRFZ44 Mosfet is only 0.28 ohms max when its Vgs is 10V. Then its max voltage loss is only 0.28V when the load current is 10A.
The on-resistance of an ordinary IRLZ44 Mosfet is 0.025 ohms max when its Vgs is 5V.
 

Daniel_B

Apr 30, 2021
7
Joined
Apr 30, 2021
Messages
7
I read many datasheets, but, the thing is, this minimum voltage drop was simply so fundamentally rooted in my mind that i simply associated these values with other things. For example, i thought the resistance is the ohmic part of the transistor and the 0.7 V is an addition to that. Diodes have this voltage drop, so BJTs, which also have a p-n junction, should also have that, at least that seemed logical.
But I'm really suprised that I didn't notice it all this time, because if one really thinks about potentials, it becomes clear that it is the voltage drop between the gate/base and the source/emitter and not some random potential (which is relative) at the base/gate that counts. Since I almost always put my transistors before the load, the potential at the emitter/source was higher and consequently, the voltage across GS/BE was lower.
 

Audioguru

Sep 24, 2016
3,656
Joined
Sep 24, 2016
Messages
3,656
What is "before the load"?
If an NPN transistor is a switch then the load is between its collector and the positive supply. If your load is between the emitter and ground then it is an "emitter-follower" and its emitter is about 0.7V less than its base voltage.
The datasheet says how much is its Vce sat when the base current is 1/10th the collector current.
 

Kiwi

Jan 28, 2013
471
Joined
Jan 28, 2013
Messages
471
"Before the load" to me would mean the transistor is configured as a "source", and "after the load" as a "sink".
Terms such as "emitter follower" and "open collector" always have me scratching what hair I have left to figure out what is going on.:)
 
Top