ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

[AQUAMAN]

Hi

I am looking to try to measure the internal gate resistance of a power transistor WHILE it is operating in a power converter.

Is it possible with an ESR meter? I know this is done for datasheets (sometimes) using an LCR meter, this can be seen by jedec standards or just through the annotations for frequency and Vac, but the transistor is out of circuit obviously. eg


However I want to try to measure this in circuit, while it is in operation, and to miliohm precision. Let's say the MOSFET/igbt is in a 5kHz single phase inverter, so the measurement has to be done FAST - I have 0.5ms or less to do it in. The capacitances are going to be in the 100nF and below range

I have looked at inferring the internal gate resistance from switching waveforms but there is non-linearity in the input capacitances and it introduces a whole load of other measurements and parasitics that need to be very accurate. I thought it would be possible to infer it from the peak gate current. For example perhaps while the transistor is OFF, I can apply a voltage beneath the threshold voltage of the MOSFET/IGBT and infer the internal gate resistance from the peak gate current and/or waveform. However I imagine this encounters similar problems to measuring the turnon/off.

Now I am looking back at similar methods to the LCR meter. I know that it is possible for ESR meters to measure down to microohm accuracy. Would it be possible to use this for gate resistance measurement? - we assume the MOSFET gate is a capacitor with an ESR (I've seen this assumed in many texts). The capacitances would be less than 10nF though.

I could build a specialised ESR meter myself and calibrate it, I could trigger it to measure in the middle of the MOSFETs OFF-state. If it runs at a frequency of 1mhz (or something high) and less than 1V, surely it should not affect the operation of the inverter? This wouldnt turn a MOSFET on with a threshold voltage of 3V lets say? The measurement would obviously have to be performed in microseconds in order to be done before the MOSFET turns on again.

To start with, I would just like to buy an ESR meter and measure with the transistor out of circuit, see if that works first, view the changes with temperature etc. Then perhaps I will start building my own one and move into a basic switching circuit at low frequency, low power. Then finding a way to measure perhaps once every minute or something, before moving to more frequent measurement in a single phase inverter or halfbridge set up

Do you think this is possible?

I have before made a system to measure the on-resistance of a transistor in the same time constraints, but I've never really used ESR meters before.

[budm]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

I am not I follow what you are trying to do, the Gate of the IGBT is insolated (basicaly capacitance between G and S (E) and G and D (C)), so what Gate resistance are you trying to measure, between Gate and ????.







https://cdn.badcaps-static.com/pdfs/...c20528318c.pdf

https://cdn.badcaps-static.com/pdfs/...454a7aff61.pdf

[fpliuzzi]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

I too was intrigued by what AQUAMAN was trying to accomplish with his high speed in-circuit dynamic IGBT testing. I noticed that the IGBT data sheets in his original post lists the internal gate resistance of the two example devices as 1.25 ohm for one of the devices and 5 ohms for the other (test conditions F=1MHZ, Vac 25mV).

The following pdf describes a method of measuring this internal gate resistance parameter for out of circuit devices using a LCR bridge (info starts at about 25% into the second page of the pdf).



Further discussion of AQUAMAN's progress seems to have moved the following thread...

https://www.eevblog.com/forum/projec...639/#msg264639

I hope he posts his final testing proceedure if he does indeed succeed in his quest. It would be interesting to read about.

Regards,
Frank

[tom66]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

There is a parasitic capacitance between the gate and source, and gate and drain. The equivalent charge that this capacitance represents is Qg, but this has a whole lot of non-linearities associated with it -- it varies over gate-source voltage for example. In series with this capacitance is a gate resistance.

In circuit, while operating, the best you can do is look at the peak current and peak voltage, but you will get maybe +/-5% accuracy (not resolution.) Why do you need to measure these parameters? Is going by the datasheet not good enough? Do you need to see how it changes over temperature, load current etc.?

[Longbow]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

Hi

I am looking to try to measure the internal gate resistance of a power transistor WHILE it is operating in a power converter.

As interesting as this project sounds, it would be great to know what your design goals are, and how the measurement of IGR would aid in that design. You are not confusing Rsat with IGR, are you?

Taking, for example, a brief overview of MOSFET switching theory as presented here: www.vishay.com/docs/73217/73217.pdf the IGR is a theoretical equivalent quantity. This quantity is calculated as an aid in understanding FET turn-on characteristics, given a step-function gate input signal. As such, IGR is more of an academic concept. The value of the IGR is derived rather than directly measured.

IGR will NOT assist you in finding the right device for a particular application. It will be the device characteristic curves, Rsat, and the actual switching speed at a given voltage that will get you closer to your design goal.

Keep in mind that even the most sophisticated LCR meter or impedance bridge will not tell you the individual values of lumped constants in a network.

[AQUAMAN]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

Yes I want to see how it evolves over temperature, load current, swithcing frequency etc. Essentially I do not care about the accuracy of the measurement within reason, I care about the repeatability and the resolution of the measurement. I want the highest resolution possible, but I dont mind if I measure say 4.0567ohms if the 'real' value is 1.0546 - as long as my measurement is very repeatable. This is the number 1 criteria. In this way I will be able to look at the changes over operating conditions.

The internal gate resistance is not really studied that much, there are very few papers on it.

In fact most people ignore it, often it isn't even included in the datasheet. However the influence on the behaviour of a device is different to what the external gate resistance has. External gate resistance limits the current into the gate, basically it just slows down the switching transition and it is just a lumped resistance. Internal gate resistance is distributed. Inside a power MOSFET/IGBT there are maybe 1000 smaller ones. The internal gate resistance means that the ones closest to the gate to turn on first, and there is a delay for the ones furthest away. If the internal gate resistance is large, this could cause quite non-uniform switching - all the load current can be shunted through a small section of the device close to the gate until the rest of it turns on. If you parallel 6 devices you could not only have all the current go through 1 device for an instance, but one small section of the device, creating hotspots, thermal runaway possibly etc

[AQUAMAN]

Re: ESR meter to measure internal gate resistance of mosfet/igbt to mOhm accuracy

I think one realistic way to measure it is to assume the gate is a capacitor and apply a small current pulse to it and measure the voltage drop. And do this during the off-state of the transistor.

As long as the voltage drop stays below the threshold voltage there shouldnt be a problem. I could even use a negative current to stop this happening. The issue is that the gate capacitances are maybe 50nF maximum, which means the current is going to need to be in the mA or less range and for perhaps 100ns pulse time (certainly below 1us i think) to avoid charging the capacitance and destroying the measurement. I am not sure if this is realistic. I will need very good resolution on the voltage measurement and synchronisation. Also the current source needs to be precise. Then I obviously have the parasitic inductance etc

At these time frames though I would probably be able to take the measurement a few times and get an average so that is a bonus.

I need to know how the ESR of capacitances lower than 100nF are reliably measured I think, and then see if this can be applied to the gate.