Ripple on DC line of Active load

[abicash]

Hi

I have built a DC Active Load required to generate current pulses ,
(please check attached schematic)

Linear FET used : IXTK90N25L2
Minimum compliance Voltage required = 85V dc
Max load = 15E ( + 2E )
Max settable current = 5A (starting at 100mA)
Pulses ranging from = 5mS to 1S.

I have a torroidal transformer which has 3 output windings 70vac,80vac and 110vac.
I am using the 80vac winding to rectify the AC through a W35 bridge rectifier and filter it with a 5000uF capacitor.
The compliance voltage thus generated is 118Vdc.

Here is the problem statement--
I get a noise of around 800mV on the ground line and the top of the square wave, across the load.
I was trying to add few X2,Y2 capacitors across the load to suppress the noise and suddenly , the top has 100Hz ticks on it on top of this 800mV , taking the actual ripple at the top to around 6Vpp (as against 800mV a day before)
I checked the capacitor voltage during switching , and it drops to 98Vdc (well above the required compliance voltage of 85Vdc)
The capacitor voltage has ripple during this switching time. which is what I believe is transferred to the load.
(although i dont know what changed , i never had this for atleast a month with the same setup)
I tried to reduce the switching time to 100mS , but i have the same noisy waveform.
I increased the capacitance to 15000uF with the same result.

Can someone please suggest if I am doing something wrong?

Thanks in advance.

 

[KlausST]

Hi,

I get a noise of around 800mV on the ground line

Usually voltages are measured with respect to GND. So how can there be 800mV GND noise?
And what value is this noise? RMS? peak-to-peak?

You know 118V x 5A gives a power of about 600W. You need a huge heatsink with fan. And you may heat your room with it in winter.

Also 5A across 1 Ohms gives 25W of power dissipation. A lot of heat.

I recommend to use a low pass filter at the signal input to avoid OPAMP input stage saturation and overshot.

Urgent: Power supply decoupling capacitors!

1F = 1As/v ... so a quick calculation with 5A, 10ms, 500uF gives a ripple of (less than) 10Vpp.

Show scope pictures with clear description of where and what you measured...

Don´t know where "square wave" comes from.
What do you mean by "during switching time"?

Don´t know what you are doing wrong, because I don´t know what you expect.

Klaus

 

[danadakk]

I dont see any caps bypassing the OpAmp. or at load supply.

Capacitor types for bypass, Polymer Tants for bulk best, and use a mix of bulk
cap and ceramic disk/MLCC.

Presumably you have a single point ground where load current and OpAmp supply meet at board
edge ?


Regards, Dana.

 

[Easy peasy]

a sketch of the WHOLE setup would be useful with arrows to exact points of measurement and their reference ....

 

[abicash]

Hi and thanks for replying.

I will try to explain my schematic and the project in detail.

The requirement is to send a constant current pulse though the load resistor (variable from 1ohm to 15ohm) programmed via a DAC. The current pulse magnitude is settable from 100mA to 5A. The resistors are all 100w each.
This is a one-time pulse, and the user may select another pulse after a while or not at all.
The pulse width is also settable from 5mS to 1S.
When the user triggers the current, it flows through the load as shown in the original schematic which i shared.

The MOSFET IXTK90N25L2 , has a linear power dissipation profile and the power which I demand lies below the SOA (from the datasheet)
I have a big heatsink and fan cooling on the MOSFET. The maximum time the MOSFET is turned on is 1sec and the duty cycle is low, as demanded by the user , which i have stated above.
The opamps are appropriately bypassed , just not shown on the schematic. (it's a concept schematic)

The compliance voltage is created by a bridge rectifier and a 5000uf/250vdc capacitor and is very smooth as seen on the DSO.

As stated earlier by me , i have written about the issues i am facing. Kindly read it above.

Note: I think there could be a common mode noise on the load resistor and I have put few capacitors across the resistor and some improvement on the noise is seen.

Kindly guide.

 

[KlausST]

Hi,

if you want me to help, answer my questions of post#2.

Klaus

 

[abicash]

Hi,


Usually voltages are measured with respect to GND. So how can there be 800mV GND noise?
And what value is this noise? RMS? peak-to-peak?

This measurement is across the load resistor at the drain of the MOSFET.
So my DSO ground line treats the voltage at the drain as ground for measurement.

You know 118V x 5A gives a power of about 600W. You need a huge heatsink with fan. And you may heat your room with it in winter.

It's for 1 second , and my actual drop is less at 98vdc and distributed across 3 resistors

Also 5A across 1 Ohms gives 25W of power dissipation. A lot of heat.

I recommend to use a low pass filter at the signal input to avoid OPAMP input stage saturation and overshot.

I will try this

Urgent: Power supply decoupling capacitors!

1F = 1As/v ... so a quick calculation with 5A, 10ms, 500uF gives a ripple of (less than) 10Vpp.

Show scope pictures with clear description of where and what you measured...

Unfortunately I am unable to get DSO pictures right now. I will share as i can.

Don´t know where "square wave" comes from.
What do you mean by "during switching time"?

Please read my earlier post.

Don´t know what you are doing wrong, because I don´t know what you expect.

Klaus

--- Updated Jun 19, 2023 ---

Apologies for not putting it up in proper syntax , I'm typing on my phone.

Please read my replies inline of your post #2

 

[Easy peasy]

a complete and very accurate schematic sketch is needed - also showing exactly how you measure