1
\$\begingroup\$

Background:

I've a problem where I need to measure the inrush current of some parallel DC/DC converters in a system where the input voltage may vary between 360-600Vdc on start up. The inrush current is roughly estimated to peak around 500A for a short duration before settling for a max current on around 15A.

I've previously tried with a mechanical relay however that particular relay, bounces on and off before making a solid contact making me unable to make an reliable measurement. Therefore I need a solid state switch solution.

The measurement is done by a current probe , P6303 from tektronix.

Idea

Despite my non previous experience of thyristors my idea is to solve this with the use of a thyristor. I think that I in this way will get a faster and less noisy switch than the mechanical relay.

At the moment I've a MCD56-12io8B thyristor laying around,which I think I can use as followed:

enter image description here or enter image description here

I have also thought of skipping the SW in the circuit and just turn on the Voltage of Vc, directly from the bench supply and hope it doesn't exceed the Vgc max.

I know that in this configuration I will only be able to turn on the thyristor however this is not a problem. I will just turn off the supply (vin) for that.

I guess the low-side solution would be easiest and safest. However I'm a bit unsure what will be the maximum allowable gate voltage for the thyristor. I can't find Vgc max or any equivalent in the datasheet to what "vgs max" is called for mosfets, which I'm familiar with, but I guess if i'm below 4V it will be chill. If you look at the pic below from the datasheet fig 5

enter image description here

Right now I would set the gate voltage to slightly above the gate trigger voltage, and the current through the gate to a value above 100mAenter image description here.

May this be a good solution to my problem or should I just buy a shit expensive ready made solid state solution..

\$\endgroup\$
  • \$\begingroup\$ Reading Fig 5 I don't think you need to worry about the gate voltage. It's determined by the slope resistance, a bit like a diode. Supply the gate from something that can supply a slightly higher voltage than needed (5V is comfortably above the lines on that graph) current limited to a safe trigger current (which looks like 200mA or so). The current limiting will determine the actual gate voltage. Whether you actually need a current source, or just 5V via a simple resistor (5 - 1.6)/0.2 = 17 ohms I can't say, hence comment not answer. \$\endgroup\$ – Brian Drummond Jan 10 '18 at 15:06
0
\$\begingroup\$

SCR are just twin transistors (PNP,NPN) such that the gate threshold is 2 diode drops until they self bias and latch. Thus the required input current is dependent on the current gain of 2 cascaded stages load and voltage drop from 2 Vbe's to the Vdc bias.

Triacs are just two SCR's for bipolar gating. Both SCR's and Triacs are defined by input trigger currents at rated loads for each polarity quadrant.

I would go with your lower schematic ( low side switch) which is easier to implement.

After successful surge testing if the DC is isolated out but may be earth grounded, then try that, which is far more stressful for leakage testing than a HIPOT test with floating secondary and relies heavily on design quality for stray capacitance of the primary AC side relative to the high f coupling impedance of the isolation transformer (if any). But if the PSU fails, you have my clue as to why.

If you wish to avoid inrush by design, then consider a 15A inrush current limiter ICL with a timer controlled bypass relay.

\$\endgroup\$
0
\$\begingroup\$

However I'm a bit unsure what will be the maximum allowable gate voltage for the thyristor.

I can see you are struggling to understand the data sheet so, maximum gate power in the table is specifed three ways: -

  • \$P_{GAV}\$ is an average and is a maximum value of 0.5 watts
  • \$P_{GM}\$ for 30 us has a maximum value of 10 watts
  • \$P_{GM}\$ for 300 us has a maximum value of 5 watts

Because you are using the gate once every several seconds I believe you should use the 5 watt figure to be a little conservative.

Looking at figure 5 and position 5 (5 watts) it suggests using a gate current of 6 amps and this results in a gate voltage of about 0.9 volts (hence about 5 watts).

Although the tables show a gate voltage of about 1.5 volts these, I believe are absolute maximums. The gate trigger currents shown just below those figures I believe should be shown as minimum values thus indicating they must be exceeded for guaranteed operation of the device at a certain speed. Remember also that the speed of the device is going to be faster at higher gate currents.

Also take note of the transient thermal impedance curves to see if you need any extra heatsinking. These demonstrate that you can rely on the thermal mass of the device for soaking up a single pulse of 100 ms duration and the temperature will rise at 0.24 degC per watt dissipated worst case. If the inrush current is over quicker then that makes heatsinking less problematic but, do take care of the steady state power dissipation because the thermal resistance is 0.6 degC per watt and if the device is operated for more than a second then heatsinking is likely to be needed but it all depends on the steady state current that flows.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.