Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

High side switched capacitive discharge

Status
Not open for further replies.

LokisMischief

Newbie level 4
Joined
Aug 23, 2017
Messages
5
Helped
1
Reputation
2
Reaction score
1
Trophy points
3
Activity points
68
Hi all,
Long time lurker here..

I am working on a high side switched discharge circuit and am struggling a little so I will start from scratch here.
What I am looking to do is to deliver a minimum of 2A through a variable resistance, lets say 2Ω to 10 or, ideally, 20Ω for 15-20ms. This only has to happen once within a reasonable time frame, so there could be serveral seconds or tens of seconds to recharge. Ideally I would like a little more current than that however, 4A would be nicer. Voltage hasnt been mentioned because current was more important however a minimum of 12v should probably be stipulated, and in any case vir on the above at 10ohm and 2A would give 20v so we should already be there.
Now, this needs to be high side switched, or even both high and low concurrently, and triggered by TTL so +5v is reasonable.

Given all this, the following was the best I could come up with, does anybody have any input or suggestions on perhaps a much better solution? I also plan to have a number of these circuits on the same pcb or atleast in the same project box, so is there maybe a better way with a common cap bank?

My Attempt
 

Cheers Klaus,

* you need a current limiting resistor between TTL input an base of bjt
* the 20V zener is not suitable, because V_GS_absoulteMaximum is 20V --> use a lower value one.
* you need a current limiting resistor between collector of bjt and gate of mosfet.

I wasnt that far off then, is this better?

* you don´t need the base-emitter resistor, but it doesn´t harm
I don't but I wanted to pull it low incase of uncertanty or partially charged gate during the startup on a µController.

* in case the load is very low ohmic you will blow the mosfet.
Was this in reference to the zenner sizing or just a general reference? The load is what I would call low ohmic in any case, do I need to alter the design here or do something to protect the FET. There is a possiblility of a short on the load also.

* the output current is not constant. It varies in time and magnitude with load impedance. But maybe you don´t want it to be constant.
* you can´t draw all of the capacitor´s energy. Only down to 3..4V. But maybe you don´t need this.
* the 1k resistor needs to be rated for at least 2.5W
The current does not need to be constant but does need to be above the 2/4A region for the 20ms

Why not low side switching?

Its actually a saftey requirement, we cannot risk exccess current to the load until we wish it, and there is a possibility that between the load and circuit could be grounded (what is not shown is a connector between the load, FET and ground) if low side switched the load is potentially sat at +V compared to the environment. This could also be us being paranoid!

Does this sound like the best solution to our problem, or is there some weird circuit or component I haven't heard of that will suit our needs?
 

edit: sorry for the wrong order of posts..
No problems, I appreciate the help.

* the 1k resistor needs to be rated for at least 2.5W
Forgot to mention, that was just there to limit the current, this will prboably be the output from a boost converter or so, so a limited charge current for the cap.


If you had a gate, then yes, but it is a bjt with base..
But as said: it will not harm...
Ahh! I guess I have forgotten much since school!

It has nothing to do with zener.
You need to protect the MOSFET according it´s SOA

Imagine the load is 1mOhms...How much current do you expect?
about 50kA, point taken. What would be the best way to protect the FET from shorts/over current then? If there is a 5Ohm resistor in series with the load that would limit the current but would also take me below the desired current pulse...

You know the discharge curve of an RC?

2200uF, 2Ohms:
peak current: 25A
time constant = 4.4ms. During this time the current drops to about 37% of the initial current.
I have tried, it would be nice to limit the discharge rate to the desired current but the variable resistance is troublesome, its one of the issues I have been having, hence me wondering if there is a better way to do this.

Safety:
I don´t think that high side switching is more safe than low side switching.

From an electrical point of view I would agree, but the load are igniter(s), I don't want the posibility that the wires from the load to the circuit could (and may accidentally) go to ground and cause premature ignition.


Loki
 

Hi,

edit: sorry for the wrong order of posts..

some issues:
* you need a current limiting resistor between TTL input an base of bjt
* you don´t need the base-emitter resistor, but it doesn´t harm
* you need a current limiting resistor between collector of bjt and gate of mosfet.
* the 20V zener is not suitable, because V_GS_absoulteMaximum is 20V --> use a lower value one.
* in case the load is very low ohmic you will blow the mosfet.
* the output current is not constant. It varies in time and magnitude with load impedance. But maybe you don´t want it to be constant.
* you can´t draw all of the capacitor´s energy. Only down to 3..4V. But maybe you don´t need this.
* the 1k resistor needs to be rated for at least 2.5W

Why not low side switching?

Klaus

*************************************
Added:

I wanted to pull it low incase of uncertanty or partially charged gate
If you had a gate, then yes, but it is a bjt with base..
But as said: it will not harm...

Was this in reference to the zenner sizing or just a general reference? The load is what I would call low ohmic in any case, do I need to alter the design here or do something to protect the FET. There is a possiblility of a short on the load also.
It has nothing to do with zener.
You need to protect the MOSFET according it´s SOA

Imagine the load is 1mOhms...How much current do you expect?

The current does not need to be constant but does need to be above the 2/4A region for the 20ms
You know the discharge curve of an RC?

2200uF, 2Ohms:
peak current: 25A
time constant = 4.4ms. During this time the current drops to about 37% of the initial current.

Safety:
I don´t think that high side switching is more safe than low side switching.

Klaus

- - - Updated - - -

Current limit:

use a 1Ohms, 2W (pulse rating is much higher) resistor in the source line,
and use a 10V zener instead of your 15V (20V) zener.

This should limit the current to about 3..4A.

Klaus
 
If you don't switch on the MOSFET fully, it will act as a current limiter if you have a fixed gate-source voltage.
It is not a high precision current limiter since there are variations with temperature and individual MOSFET's.
Assume that the variation is a factor 2, so if you aim for 4A the current limit should be 2A - 8A. This means that the MOSFET will have a very low voltage drop as long as the current is below 2A.
With this current limit, you can easily find a MOSFET that will survive a dead short for 20 ms. Just look at the SOA (safe operating area) diagram in the data sheet.
Look for 8A and the supply voltage, and find a MOSFET that can handle that for 20 ms.

With the resistor in the source line suggested by KlausST, the current limit will have a higher precision but the drawback is the very high voltage drop. The supply voltage must be much higher for the same voltage/current output.
 

Current limit:

use a 1Ohms, 2W (pulse rating is much higher) resistor in the source line,
and use a 10V zener instead of your 15V (20V) zener.

This should limit the current to about 3..4A.

Thanks Klaus, this seems to be doing the trick though I dont have any 10v zenners on hand, only 15v, I presume this was to limit the Vgs but given the Vgs limit for that perticular FET is 20v I dont understand why 15v isn't suitable? I have probably missed something!

std_match, Is this not sort of what we are doing with the zenner here? Given the source voltage changes with time (as the cap discharges) how would you suggest achiving your suggestion?



Now for the next challange. (and tell me if this is better suited to a new thread or if more suited to the digi section.)

I would like to know what the resistance of the load (probably 1-30ohm range) is, generally speaking I would just create a voltage devider and measure the voltage drop with a reference voltage, however I need to keep the test current at ideally 5ma or less.

So, I am out of ideas here, does anybody have any suggestions on both multi channel ADC's (8 or 16 ideally if not more) and any suggestions on how to measure the load resistance while keeping current around or under 5ma, taking into account the fact there will be +50v passing through the load at some point also so probably a protection diode too. As for accuracy, whole ohms are acceptable, this is more a guide than anything else.
In the worst case I would accept continuity (which I could probably manage with +5v, 10k resistor and a diode) but would rather know the resistance.

Loki
 
Last edited:

Hi,

I dont understand why 15v isn't suitable?
10V is for the current limitng feature.
With a 15V zener in this case you limit the current to about 6..9A.
****

Measure load resistance:
Build a precise 5mA current source.
Now you want to connect a load in the range of 2Ohms ... 20 Ohms.
This means the measurement range needs to focus on 10mV ... 100mV.

I recommend a filter, an amplifier with gain of 10 and voltage limiting capability.

Feed the signal to an ADC.

***
The more simple solution is an amplifier with gain of 2 and feed the signal to a 0..200mV ready to buy voltmeter.

Klaus
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top