LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

characteristic defect

Solved!
Go to solution

The BS170 you specified has a spec sheet here - https://www.onsemi.com/pub/Collateral/BS170-D.PDF

 

>> I want to add a resistor to the drain so I can change the voltage till 10V without burning my transistor. 

 

Putting a large resistor in the drain means you limit the voltage at the device. Its a simple voltage divider.  Rds is ~ 2Ohms from the spec sheet, so the voltage divider is Vresistor = 10V*1000/1000+2 = 9.98V across the resistor and Vds = 0.02V.  So the current you measure is the same current you would get if you bias your transistor with Vds = 0.02V.   Easy to verify!

 

So you don't want to do that. Spec sheet says the Max Vds = +60V, so you should be fine at +10V.   The DMM will since up to 2A, so you are fine running without it.

 

Yes, the transistor will warm up.  Some of mine operate at 400-500C. Smiley Happy

 

>> Ramp by pattern 

Its in the Signal Pallette, but easiest to use Quick Drop (Ctrl + space) then search "ramp".  

 

Craig

Message 11 of 27
(183 Views)

yeah you are right, from the datasheet I should be fine at +10V, but I've just tried without resistor to 9V and my transistor burn out .

In the plot you will see that at VG=6V ,I get to the breakdown region, wich is not expected.test.PNG

0 Kudos
Message 12 of 27
(179 Views)

I've just find out that variable power supply Max output:12VDC, 500mA.

DC accuracy for DMM :range 2A= 1350ppm+180ppm .

0 Kudos
Message 13 of 27
(173 Views)

Typically you do not use a DMM and PSU to characterize a transistor for this sort of testing, ideally you use SMUs which were designed specifically for the purpose.  The problem is that the devices aren't stabablized and any spike in Vgs can cause problems - PSUs can be too slow and noisy. 

 

To compensate you should add a bias tee or capacitors in both gate and drain to reduce spurious effects when the PSU changes values.  It will change your characteristics to a small degree but make the testing more stable.  The other options are to ramp the voltage more slowly, add shielding to the setup, use shielded cables, etc..

 

Craig

Message 14 of 27
(171 Views)

I tried again with a step of 0.25 for each VG and Vds, and my transistor did not burn out with 9V, but from the characteristic my transistor is not stable. I want to add capacitors to the drain and the gate but I don't know how to choise the value of the capacitors.TEST2.PNG

0 Kudos
Message 15 of 27
(150 Views)

hi

 

The problem is that the devices aren't stabablized and any spike in Vgs can cause problems - PSUs can be too slow and noisy.

 

I am  still trying to solve the problem, can I apply a function that block the HF or the parasites  before the acquisition of my signal in the VI? 

 

 

0 Kudos
Message 16 of 27
(123 Views)

Yes.  You should look at how bias tees work - https://en.wikipedia.org/wiki/Bias_tee

 

Your PSU is the RF and DC source.  You want the inductor in line with your device and the capacitor terminated to 50Ohms.  Now the values of L and C depending on knowing the characteristics of your source of noise and what you need to filter out.  Have a look at the ready made solutions available for a few bucks here..

 

https://www.minicircuits.com/WebStore/BiasTees.html

 

But have you tried just using delays before measuring?  Coaxial cables and removing the breadboard will also help.

 

Craig

0 Kudos
Message 17 of 27
(119 Views)

I checked and I find this, how should i choose the delay,  :

slew rate of the output is 20V/us.

timing resolution=50ns.

settling time of the input is for range of 10V :10LSB--1us, 1LSB--2us.(LSB?)

timing resolution=50ns.VI IV curve tracer.PNG

I tried with a delay of 1s and a step of voltages 0.25 and i get this :

TEST2.PNG

for the parasites I checked and I find this:

PSU : ripple and noise 25mV.

AI absolute accuracy: noise full scale 280 uVrms.

 

0 Kudos
Message 18 of 27
(107 Views)

Coaxial cables and removing the breadboard will also help.

but how I can remove the breadbord because I am using NI Elvis breadbord .

0 Kudos
Message 19 of 27
(104 Views)

Put your transistor in a shielded metal box.  Use BNC feedthroughs, connected to BNC cable to the NI-Elvis board.  At the very least add stabilizing circuits to your bread board setup.

 

I'm not going to do your homework for you, but try 100pf and 50nH.  Learn the bias tee theory and adjust as required.

 

Craig

 

 

Message 20 of 27
(101 Views)