Dummy Load with Mosfet

Status
Not open for further replies.

hamid159

Full Member level 3
Joined
Aug 25, 2013
Messages
176
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,298
Location
Lahore
Activity points
2,488
Hi guyz,
I'm making dummy load using mosfets ( about 200Watt and 5 - 150V )
But there is a problem whenever I change the voltage let say 5v the maximum current should be drawn from the source ( to be checked) is 200/5 = 40A...now I change the voltage to 100V,then the maximum current should be 200/100 = 2A..But my circuit will make the source to draw maximum 40A current...and everything will boil in a split of second ( power dissipation of 40*100 = 4000W)...How can I make this circuit not to dissipate the power greater than 200Watt?
Only one stage is shown in circuit diagam...There will be total 4 stages with each mosfet paralleled with two more mosfets..and each stage can make the source to draw maximum of 10A.So,There total maximum current will be 10A * 4 stages = 40A (at minimum voltage i.e, 5V )
 

Attachments

  • variable_voltage_dummy_load.PDF
    23.3 KB · Views: 227

Hamid you have very high gain in a design that compares current not power to the variable voltage. If you want to control dummy load power dissipation regardless of supply voltage, then you must monitor power. One way to monitor power is to mount a thermistor to the heatsink and use some anticipation or derivative feedback ( part of PID control) to regulate power. The body temperature will stabilize to a temperature above ambient that is proportional to dummy load power. This requires a thermistor bridge and PID loop. One thermistor for ambient and one mounted on the heatsink..The other way is to measure instantaneous power using an analog mulitplier for Vds and Ids with an output to compare with reference voltage. .As power increases, the control voltage must decrease to compensate for this MOSFET characteristics in the range from 6V towards 5V.
 

Thanks Sunnyskyguy for your suggestion,

I think using thermistor is not an easy task...
you said..use analog multiplier but multiplier has limited switches and I have 5 - 150V range..how I can use that to measure instantaneous power
 
Last edited:

..............................
you said..use analog multiplier but multiplier has limited switches and I have 5 - 150V range..how I can use that to measure instantaneous power
Use a resistive divider to reduce the 150V to the maximum voltage the multiplier input can tolerate. Add a small resistor in series with all the MOSFET sources to monitor the current.
 

Hi,

what you built is an adjustable constant current sink.
with increasing voltage the current is constant. But the power is linear to voltage.
Compared to a resistor as load you are on the half way to your goal.

If you are familiar with uControllers then use it to measure the voltage and calculate the desired current.
This is the setpoint to U1B.

I recomend to use a local HF feedback for U1B to avoid oscillatiing.

For the use with your 10 mOhms shunt i recomend to use a low offset opamp.

Good luck
Klaus
 

Thanks crutschow,,
But series resistor with the mosfet sources has already been added as shown in circuit diagram...Sorry,,i misunderstood the concept of multiplier with multiplexer ...Can you explain how muliplier will work in this circuit .i.e, how it will examine the power and not the current???
And Thanks to Klausst
,,,Yes i'm familiat with uControllers...if i measure the voltge,and calculate the desired current..then how to go further,if microcontroller can provide variable voltage,then it should be very easy..but problem is that how microcontroller will work in this situation..
 

So you want to build a load with some sort of power related control that will limit the total power to 200W. As you do not need to measure the power or have it very linear I am trying to think of a cheap and nasty solution. If you can derive Vv which is the LOG of the output volts and Vi which is the log of the output current, then adding them and passing them through an anti log amplifier will give the result, V X I = power, but if you do your comparison with log 200 (W), then you do not need an anti log amp.
So first of all equalise your V and I levels, so 150V goes through a 30:1 attenuator. Your current sensor (20mV @2A) goes through a 250 X amplifier, so both giving 5v at max level. Feed both these signals each via 6 X Si diodes into a current summer (100 ohm resistor), compare the voltage drop across this resistor with a fixed DC with an opamp, which removes (or reduces) the FET drive and lights a red warning LED.
So if it works, 12 diodes a few resistors and an op amp. (IF it works :-?)
Frank
 

Sorry I didn't take a close look at you circuit to see the source resistor

If you multiply the voltage from the op amp measuring the current through the MOSFET times the voltage across the MOSFET then the multiplier output voltage represents the power dissipated in the MOSFET.
 

Chuckey,i don't want to turn off my circuit..if I have to,then it should be so simple.

- - - Updated - - -


Analog multiplier has four inputs,,inverting1,non-inverting1,and inverting2,non-inverting2..how should I connect them with voltage and current?
 

uC can multiply result, output to DAC and then to Op Amp which compares desired power level.

Here is a variation of what I suggested with two thermistors, except instead of large heat mass of MOSFET, using small resistor heater thermally bonded to thermistors ( such as with Bondex uses high metal particles in epoxy)

Then your voltage v1,v2 are scaled to say 5V range from shunt and V max range.

**broken link removed**

For a faster time response, supply a tiny fan for fast linear air speed.

The same could be done with MOSFET as heat source, THe thermal resistance is lowered with the fan, providing more dummy load capacity and still acting as a power sensor, making a simple dummy load control system. Calibration can be done with meters to measure time error and improve feedback to reduce error.
 


Pretty Cool...But a lot of components are required in this case......if analog multiplier can do the job.i can use that,if you guide me how to connect it.
 
Last edited:

Hi,

Ucontroller solution:
measure voltage, calculate current. Power / voltage = current.

If you want it fast, then output it with a DAC,
More slowly, but more simple is to output the current with pwm and filter it.

Feed the current signal from DAC/PWM to your already existing current regulator.

Hope that helps
Klaus
 

............................
Analog multiplier has four inputs,,inverting1,non-inverting1,and inverting2,non-inverting2..how should I connect them with voltage and current?
The multiplier has differential inputs. If the voltage and current voltages are both positive, then ground the inverting inputs and connect the voltage and current signals to the non-inverting inputs.
 

Hi Klausst,
If my voltage is constant let say 20V..then the output from the uC will also be constant...then if i turn my variable resistor to change the current then the input to uC is constant i.e,20V..... and i don't want my power to be constant at 200W..this is just maximum power that can be dissipated...
 

Hi,

... then i misunderstood ...

Can you give a formula about voltage and current?

Or a table with your desired values? something like that:

V; A
5; 40
10; 20
20; 10

Maybe you want something like that: (I´d call it a constant current sorce (10A) with power limit (200W))
V; A
5; 10
10; 10
20; 10
25; 8
40; 5
50; 4

Klaus
 

yes...table should be like that...but voltage can vary even 0.1V,,,and current should be adjusted from variable resistor but maximum current should be = 200W/voltage....
V; maximum A
5; 40
5.1; 39.2
5.2; 38.46
.
.
.
.150; 1.33

but this is the maximum current...any value of current lower than maximum should be adjusted by variable resistor..

by the way,This dummy load will check the power supply...e.g,i make a smps..and can check it with dummy load from no load to full load.
 

Hi,

but this is the maximum current...any value of current lower than maximum should be adjusted by variable resistor..

My ideas of testing an SMPS:

1) connect SMPS to TestLoad
2) setup TestLoad to draw 0A
3) switch on SMPS
4) setup TestLoad to draw 1A (or any other value)
5) Check if SMPS is correctly working
6) setup TestLoad to draw more power (+ 1A ??)
7) repeat step 5 and 6 until SMPS function fails
8) switch everything off

Is this what you want?

Now you have one problem to solve: adjusting current with resistor and adjusting current with the power limit function with ucontroller.
With the potentiometer connected like in your circuit this is not easy....

Therfore i have two ideas for you.
Deciding what current value (potentiometer or powerlimit) is smaller is an easy task for a ucontroller. So you can connect the potentiometer to another ADC of your ucontroller.
- The controller reads both ADC values: Applied voltage, and potentiometer.
- With the applied voltageit calculates the max. current
- Now it takes the lower of the two values and outputs it to the U1B, pin 5. (with DAC or filtered PWM)

finished....

but why not do the next step and control it with your PC.
- communicate with the PC with UART, USB, bluetooth or anything else.
- the PC transmits the desired current to the controller
- and the controller transmits the voltage to the PC
- You can use VB, C, VBVA in Excel to automatically test your SMPS, you get a complete U-I chart. or long time tests....

A useful feature is to connect a comparator to pin7 of U1B:
A threshold near upper rail tells the ucontroller that the current drawn is too low. Maybe because of too low SMPS voltage.
Transmit this information to the PC, too.



Hope this helps
Klaus
 

Hamid, it is unclear what you want to accomplish with this design. "and i don't want my power to be constant at 200W..this is just maximum power that can be dissipated.."200W will require a large heatsink and high linear velocity airflow even if shared among 4 FETs.I still think the thermistor feedback is the best solution, to not only regulate your demand power with temperature rise measurement , as well as overtemp protection.
 


what do you mean by filtered pwm?
i don't know much about communication uC with pc...
 

Hi,

filtered PWM: PWM is a digital signal with varying duty cycle. Duty cycle can be 0....100%. With an analog low pass filter you can generate a varying DC voltage. From 0% to 100% VCC.

Communication with PC: There is no need to communicate with the PC. Then just connect the potentiometer (wiper) to an ADC input. One side to GND, the other to VCC.
If you have a 10 bit ADC then you get values from 0...1023 (ideally).
Decide what current this should mean. Maybe 0...20A.

For measuring voltage you need a voltage devider. A good start is to select the lower to 10k Ohms. Select the upper to fit your max. SMPS voltage and your VCC.
R_min = 10k * (V_SMPS / VCC -1)

Good luck

Klaus
 

Status
Not open for further replies.

Similar threads

Cookies are required to use this site. You must accept them to continue using the site. Learn more…