Engineering Analyzing the switch circuitry for an HF Amplifier

[ms002]

Homework Statement

I am trying to figure out what every transistor in this circuit does. how do they function together and what kind of effect should I expect

Relevant Equations

Q1, Q3, Q4, 6 are NPN Transistors
Q2 is a p mosfet
Q5a is n mosfet

 

[Baluncore]

Welcome to PF.

That looks like a LTspice schematic, but it is hard to read.
Please attach a copy of the schematic file to your next post, so I can read and run the simulation. Extend or change the file extension to schematic.asc.txt so it can be attached to the post, since it is an ascii text file on the inside.

 

[ms002]

Welcome to PF.

That looks like a LTspice schematic, but it is hard to read.
Please attach a copy of the schematic file to your next post, so I can read and run the simulation. Extend or change the file extension to schematic.asc.txt so it can be attached to the post, since it is an ascii text file on the inside.

iam sorry i forgot to attach the file. here you go. and thanks in advance

 

[berkeman]

Welcome to PF.

Since this is your schoolwork problem, you need to tell us what you think before we can provide tutorial help. What are your thoughts, and what have your simulations so far shown you? Please attach screenshots of your simulations so we can better understand your work so far. Thank you.

 

[Baluncore]

iam sorry i forgot to attach the file. here you go. and thanks in advance

For which BJT transistors, do you recognise the circuit and purpose?

Where does the circuit come from? It looks like it was designed by a committee.
It has a DC gain of about ten. If you have run the simulation, then verify that it works with an 8 volt input signal, rather than the 40 volt V2 input overdrive.

Can you give me a web link to the component library zip file you use, rather than the embedded link to your HDD. ".lib ????? \Github\Amplifier\libs\Bib.LIB"

Avoid decimal points with LTspice, use a single letter.
0.1u = 0u1; 4700 = 4k7; 220 ohms = 220R; 2.2 ohms = 2R2

R3 to R12, with C2 to C11, are connected across a pure voltage source. They play no part in this simulation, but will play a part in the physical realisation

 

[ms002]

For which BJT transistors, do you recognise the circuit and purpose?

Where does the circuit come from? It looks like it was designed by a committee.
It has a DC gain of about ten. If you have run the simulation, then verify that it works with an 8 volt input signal, rather than the 40 volt V2 input overdrive.

Can you give me a web link to the component library zip file you use, rather than the embedded link to your HDD. ".lib ????? \Github\Amplifier\libs\Bib.LIB"

Avoid decimal points with LTspice, use a single letter.
0.1u = 0u1; 4700 = 4k7; 220 ohms = 220R; 2.2 ohms = 2R2

R3 to R12, with C2 to C11, are connected across a pure voltage source. They play no part in this simulation, but will play a part in the physical realisation

the circuit is designed as a switch for a hf amplifier. iam trying to recreate the switch itself. i know that Q2 and Q5a activates and deactivate depending on whether the Voltage is positive or negative. push pull function. Q3 and Q 4 are differential amplifier. iam not sure why are they there tho. Q6 grounds the signal when the voltage is negative. iam not sure about the function of Q1
i know that R2 and R19 is a voltage devider. and R 16, R20, C13 is a low pass filter.
iam not sure what R58, R59 and R22 do but iam assuming its something to do with Vp.
My goal is to simulate the functionality of a switch and to understand what each transistor do.
so Vamp should be 0 what V2 is low( 0 ) and 56 when V2 is high (40). doent matter to me what is the high number. it could also be 5. its just a random pulse i created to reconstruct a button

 

[ms002]

Welcome to PF.

Since this is your schoolwork problem, you need to tell us what you think before we can provide tutorial help. What are your thoughts, and what have your simulations so far shown you? Please attach screenshots of your simulations so we can better understand your work so far. Thank you.

hello, sorry for the trouble. i posted a summary of how far i got as a reply

 

[dlgoff]

i posted a summary of how far i got as a reply

I don't see that.

 

[TonyStewart]

Can you see the problems of a complementary drain linear feedback to try to feed> 1kW from FET power switches using RC negative feedback effects on gate drive? In terms of latency, slew rate , optimal bias and suppressed gate swing, lack of dead-time?

This demands extremely high dV/dt= Ic/C with extremely high surge currents in parasitic and fixed capacitors.

A far better approach recognize the effects on all capacitors and the ones not shown for parasitic and Ciss , Coss of FETs and C13. Also realize the time constants on real caps are wrongly shown beside it is using an ideal voltage source.

Examine BLDC and Piezo drivers with Deadtime control instead. Not this.

Modified to work. But ignoring overcurrent failures.

 

[ms002]

Can you see the problems of a complementary drain linear feedback to try to feed> 1kW from FET power switches using RC negative feedback effects on gate drive? In terms of latency, slew rate , optimal bias and suppressed gate swing, lack of dead-time?

This demands extremely high dV/dt= Ic/C with extremely high surge currents in parasitic and fixed capacitors.

A far better approach recognize the effects on all capacitors and the ones not shown for parasitic and Ciss , Coss of FETs and C13. Also realize the time constants on real caps are wrongly shown beside it is using an ideal voltage source.

Examine BLDC and Piezo drivers with Deadtime control instead. Not this.

View attachment 328369Modified to work. But ignoring overcurrent failures.

thaank you. that was most helpful. i will take alook at the sources you mentioned.
i see how faulty it is and not really any good for practical use. it was designed by a master student so i didnt expect it to work that well. i dont think i would be using the schematices anyways. It's just fun to try and experiment with different kinds of circuits. for me its more about understanding the functionality of each transistor. what i mean is, what does each transistor do. i kind of have a general idea as mentioned in my earlier reply. iam just not sure if my understanding is correct. its just for me to better understand different kinds of transistors. the different effects that each transistor has in this simulation and how do they all function with eachother

 

[ms002]

I don't see that.

its two replies above :)

 

[Baluncore]

I am not sure about the function of Q1 ...

Q1, D1, R2 is an active pull-up circuit, designed to turn the MOSFET off faster than the other MOSFET turns on. When Q4 sinks current, it turns on the p-chan, upper MOSFET by sinking current through the diode. When Q4 does not sink current, Q1 actively pulls the gate voltage high, turning Q2 off quickly. During a transition, the capacitance of the Q2 gate, is charged by Q1, or discharged by Q4.

... and R 16, R20, C13 is a low pass filter.

R16 and R20 set the gain of the amplifier to about 10, by attenuating the feedback signal. C13 is not really a low-pass, it is a high-pass that speeds up the feedback for high frequencies, probably to improve stability.

I know that R2 and R19 is a voltage devider.

Current through R19 is set by Q4 base and emitter voltage, that current flows through R2, providing or dropping sufficient voltage to turn Q2 on. The input voltage only needs to be about 8 volts, since that will sink enough current through R19 to provide enough gate voltage across R2 to turn on the P-chan Q2.

 

[ms002]

Q1, D1, R2 is an active pull-up circuit, designed to turn the MOSFET off faster than the other MOSFET turns on. When Q4 sinks current, it turns on the p-chan, upper MOSFET by sinking current through the diode. When Q4 does not sink current, Q1 actively pulls the gate voltage high, turning Q2 off quickly. During a transition, the capacitance of the Q2 gate, is charged by Q1, or discharged by Q4.R16 and R20 set the gain of the amplifier to about 10, by attenuating the feedback signal. C13 is not really a low-pass, it is a high-pass that speeds up the feedback for high frequencies, probably to improve stability.Current through R19 is set by Q4 base and emitter voltage, that current flows through R2, providing or dropping sufficient voltage to turn Q2 on. The input voltage only needs to be about 8 volts, since that will sink enough current through R19 to provide enough gate voltage across R2 to turn on the P-chan Q2.

this is most helpful to understand the functionality of the circuit. I am extremely grateful. thank you

 

[DaveE]

R16 and R20 set the gain of the amplifier to about 10, by attenuating the feedback signal. C13 is not really a low-pass, it is a high-pass that speeds up the feedback for high frequencies, probably to improve stability.

But why have negative feedback in a switch? I don't understand that part.

 

[Baluncore]

But why have negative feedback in a switch? I don't understand that part.

It appears to have grown from a legacy of three or more people's experience, without understanding. The circuit was designed by a committee, until it worked well enough. As artists, any one of us would do it differently today, but it is what it is, so to avoid confusion I will work with it.

 

[ms002]

Q1, D1, R2 is an active pull-up circuit, designed to turn the MOSFET off faster than the other MOSFET turns on. When Q4 sinks current, it turns on the p-chan, upper MOSFET by sinking current through the diode. When Q4 does not sink current, Q1 actively pulls the gate voltage high, turning Q2 off quickly. During a transition, the capacitance of the Q2 gate, is charged by Q1, or discharged by Q4.R16 and R20 set the gain of the amplifier to about 10, by attenuating the feedback signal. C13 is not really a low-pass, it is a high-pass that speeds up the feedback for high frequencies, probably to improve stability.Current through R19 is set by Q4 base and emitter voltage, that current flows through R2, providing or dropping sufficient voltage to turn Q2 on. The input voltage only needs to be about 8 volts, since that will sink enough current through R19 to provide enough gate voltage across R2 to turn on the P-chan Q2.

what role Q3 plays in this circuit. i cant quite figure out what it does?

 

[Baluncore]

Q3 is an artefact of the MOSFET (audio or servo) power amplifier that seeded the redesign process. Q3 and Q4 are no longer needed as an error amplifier, since it morphed into a digital switch.

The control input now overdrives the analog amplifier to saturation, which makes it digital.

There are some problems with the existing circuit because both MOSFETs are on at the same time during the turn-on transition. It is only for 50 nsec, but the through-current peaks at 16 amps.

1. Why is this analysis of a problematic circuit in the homework forum?
2. Is this a communications or radar application, or an enclosed instrumentation application such as NMR?
3. How often, and how quickly, must the power supply to the HF amplifier be turned on or off?

 

[DaveE]

what role Q3 plays in this circuit. i cant quite figure out what it does?

It provides a path for negative feedback. It is part of a differential pair with Q4. When it's emitter current increases that increases the voltage across R19, which is also the emitter voltage for the Q4. Increasing the emitter voltage on Q4 will act to decrease it's current.

This differential amplifier circuit is an important one to learn about.

OTOH, I have no idea why they wanted that negative feedback in this application. I would have left it out, I think.

 

[ms002]

Q3 is an artefact of the MOSFET (audio or servo) power amplifier that seeded the redesign process. Q3 and Q4 are no longer needed as an error amplifier, since it morphed into a digital switch.

The control input now overdrives the analog amplifier to saturation, which makes it digital.

There are some problems with the existing circuit because both MOSFETs are on at the same time during the turn-on transition. It is only for 50 nsec, but the through-current peaks at 16 amps.

1. Why is this analysis of a problematic circuit in the homework forum?
2. Is this a communications or radar application, or an enclosed instrumentation application such as NMR?
3. How often, and how quickly, must the power supply to the HF amplifier be turned on or off?

thank you for ur answer. understandable. its not quite clear why did they build it the way they did.
to answer your questions. i just needed help understanding how this circuit works. i found it interesting ow it was constructed.
its for a MRI. as a signal amplifier.
thas the thing. in reality, it's not a set amount of times. it should work like how a normal amplifier works for an MRT. i created the pluse to simulate an automated switch because LTspice doesnt offer that as an option

 

[ms002]

It provides a path for negative feedback. It is part of a differential pair with Q4. When it's emitter current increases that increases the voltage across R19, which is also the emitter voltage for the Q4. Increasing the emitter voltage on Q4 will act to decrease it's current.

This differential amplifier circuit is an important one to learn about.

OTOH, I have no idea why they wanted that negative feedback in this application. I would have left it out, I think.

thank you. i thought so as well. i was just confused why it was built that way. because in all modern applications, the two transistors would be replaced with a digital amplifier

 

[Baluncore]

I found it interesting how it was constructed.
It's for a MRI as a signal amplifier.

MRI = NMR imaging, but without the Cold War scary "Nuclear" word.

So the circuit is the switch that gates the power to the HF PA used for MRI excitation. That does explain why it has a pull-down MOSFET to turn it off fast.

The technician who threw the design together used what schematics were available at the time. That does explain the design, because it looks a bit like the linear power amp for MRI did 30+ years ago. I have an original MRI PA under my bench here, it makes a good footrest, too heavy to move. Electronic archaeology is always easier if you were there at the time.

My guess is that c2 to c11, and R3 to R12, were a model of the PA internal power distribution network, so would be connected to the output of the power supply switch, not to the switch input supply rail.

If you are going to build one, and would like the circuit updated, we can probably do that.
I would expect today's HF PA design would be internally gated by a logic signal, rather than switching off the entire HF PA power supply.