Gate Drive Transformer Secondary parameters Design

Thread Starter

picstudent

Joined Feb 3, 2009
91
I am trying to do a half bridge like this
HalfBridge3525.jpg
I got a fairly good idea about how to design the primary side of T1. Calculating N minimum for primary and then inductance, magnetizing current and then peak current etc. I got sufficient literature to understand the concept. The secondary I need is 1:1, so obviously I think same number of turns and many articles also describing that ratio, But noting else.
I am curious, how we can make sure that the secondaries can provide enough drive to charge the MOSFET gates and maintain safe voltages there to keep MOSFETs on.
Thanks
 

ronsimpson

Joined Oct 7, 2019
2,986
coilcraft ... More coilcraft
A transformer is not good if the duty cycle is very high or very low. It works well if the design runs 20 to 80%.
Put a scope across R2 and look at the drive.
Watch that the voltage does not get out of range +/-20V. Keep above the gate turn on voltage.
1596372730175.png
1596372754881.png
1596372781338.png
 

Thread Starter

picstudent

Joined Feb 3, 2009
91
coilcraft ... More coilcraft
A transformer is not good if the duty cycle is very high or very low. It works well if the design runs 20 to 80%.
Put a scope across R2 and look at the drive.
Watch that the voltage does not get out of range +/-20V. Keep above the gate turn on voltage.
Thank you very much for the information.
My duty cycle is average, so should not be a problem. Being half bridge SG3525 in this mode can only go up-to 50%. Correct?
I was looking for almost the same circuit for Generating high voltage.
R2 I cannot Find in the schematic, I think you are suggesting to measure Gate to Source voltage.
Thanks again
 

ronsimpson

Joined Oct 7, 2019
2,986
R2 I cannot Find in the schematic, I think you are suggesting to measure Gate to Source voltage.
In post #1. G-S voltage.

A transformer can not pass DC. (unless some games are played)
Watch out for what happens at min and max duty cycle. I am not talking about the duty cycle of the IC but of the signal on the transformer. The middle picture if of 50:50.
1596394186291.png
 

Thread Starter

picstudent

Joined Feb 3, 2009
91
This is the two outputs of IC , in each channel of the scope Referenced to Ground
Scope1.JPG
This is the signal when SG3523 output loaded with a 1K resistor and probe between two ends of the resistor
Scope2.JPG

Up-to this part seems ok.
When I connected my Gate Drive transformer, SG3525 is getting heated up and even one IC damaged.
Needless to say, the output of the transformer is a mess!

What could be wrong?. GDT primary is showing around 900uH in my inductance meter which is in agreement with my calculations.
 

ronsimpson

Joined Oct 7, 2019
2,986
Do some math: You have nearly the full supply voltage across the transformer for 20uS. What is the current? I can not read your scope. Is it 20uS positive, 1uS zero volts, 20uS negative, 1uS off?

Try adding 4 Schottky diodes like in Post #2 (schematic).
 

Thread Starter

picstudent

Joined Feb 3, 2009
91
I can not read your scope. Is it 20uS positive, 1uS zero volts, 20uS negative, 1uS off?
When I connect both outputs to each of the channels of the scope, the voltage swings from 1.6V to 14V. (Supply is 12Volts). The positive cycle is around 19uS and Negative (Off Time) is around 21uS. A small dead time of around 1uS is there. This is the first image.

Next I connected the Scope probe between two outputs.
Now the output swings between (as per the scope readings) +15V and -9.6V. Dead time can be seen there in second picture. ( This negative Voltage I cannot understand). ( Scope is Powered through a isolation transformer)

Do some math: You have nearly the full supply voltage across the transformer for 20uS. What is the current?
I have 18 turns in my torroid with Al value around 2500nH/N^2.
So L =2500*18*18=810uH
I=(12V*.5)/(2*810*10^-6*25*10^3) = 148mA (.5 is of 50% duty cycle and 25 kHz freq gives 20uS)
I think this is the peak current. (Please correct me if my calculation is wrong)
Now this is the datasheet portion about output current.
SG3525Current.jpg
Which current we need to consider in this case?

Try adding 4 Schottky diodes like in Post #2 (schematic).
Not yet added . but going to try that.

Thanks for your time.
 

ronsimpson

Joined Oct 7, 2019
2,986
Some thing is making the IC hot and I think it is too much current in the output.
Disconnect the transformer's secondary that goes to the top MOSFET.
Now connect it with the primary. Not parallel because that does nothing. Series! Watch the phase. So the inductance goes up not down! Inductance should be 4X . Does the IC run cool? (if you connect series wrong the inductance goes to zero)
 

Thread Starter

picstudent

Joined Feb 3, 2009
91
Try adding 4 Schottky diodes like in Post #2 (schematic).
I think these diodes solved my issue.

Now the wave forms are reasonably good. (Almost full Duty cycle). I hope that the top notch (slight over shoot) in the transition time is not that much a big issue. Correct?
The picture shown is the wave from in two secondaries of the GDT caught on two channels on the scope across 560 Ohms resistor.
Only problem is the slight variation voltage amplitude on two secondaries
14V and -9V in one winding and 16.8V and -12.4V in other winding even though I wound the wires in twisted combination

GDT_Correct.JPG

Thank you for the support
 

ronsimpson

Joined Oct 7, 2019
2,986
Only problem is the slight variation voltage amplitude on two secondaries
Connect both probes to the same signal. Connect in the same way. See if the signals look the same. Have you done a probe calibration? There should be a adjustment so both probes read the same a low frequencies and DC. Then another adjustment for high frequencies. Before we blame the transformer, check the probes.
 
Top