Go Down

Topic: TTL Signal from using Arduino Rev 3 to MOSFET Driver (Read 1 time) previous topic - next topic

YMute

I have been trying to make a pulse generator using Arduino Uno Rev 3 to drive MOSFET driver, which then drive MOSFET, by using TTL signal from Arduino. (5V in as Vcc and about 4.8 V in as TTL signal)

I am currently using Arduino Uno Rev 3 to supply voltage to MOSFET driver, IXDN614PI, which is connected to a MOSFET.

R2 and R3 are shunt resistors to prevent any damages from reverse flow of voltage when MOSFET and MOSFET driver is switching on and off.

However, when I run the Arduino, it starts off fine showing good square TTL signals, however as time passes, the driver no longer generate enough voltage to drive the MOSFET or even shows some signs of damages on signal output to the MOSFET.

And when I even removed the MOSFET and try to see pulse generation from MOSFET driver by oscilloscope, then it has same problem.
MOSFET driver runs for up to 30 minutes then dies down.
It no longer shows pulse at all or shows some damaging pulse, not a square, but a deformed square.

I have attached the schematic of connection that I have made for driving MOSFET driver.

May I get some suggestions on how to generate constant TTL signal from MOSFET driver?

MarkT

Perhaps a stronge pullup on the enable pin?  Is the decoupling ceramic and within 5mm of the package?

Oh, and is it getting hot?
[ I DO NOT respond to personal messages, I WILL delete them unread, use the forum please ]

JohnRob

Another possibility is too low a gate voltage.  For this mosfet the Gate threshold voltage ranges from 2.7 to 5.6 V.

The Gate Threshold Voltage (Vgs) is the voltage where the MosFet just starts to turn on.  I don't recall the temperature effect on Vgs but perhaps you simply don't have enough gate voltage.

Please do not PM me with thread based messages.  If your thoughts are worth responding,  the group should benefit from your insight.

raschemmel

#3
Oct 22, 2018, 07:46 pm Last Edit: Oct 22, 2018, 08:21 pm by raschemmel
Well ok, if we are going to troubleshoot the circuit using an oscilloscope, how about showing us the scope
screen images (before/after) by posting photos taken with your cellphone ?

Haven't seen any mention of chip temperature ? Have you not touched the driver chip when it starts to
degrade ?

Also, no mention of gate driver Power voltage (VDD . What voltage is the driver running off ?

Quote
And when I even removed the MOSFET and try to see pulse generation from MOSFET driver by oscilloscope, then it has same problem.
MOSFET driver runs for up to 30 minutes then dies down.
It no longer shows pulse at all or shows some damaging pulse, not a square, but a deformed square.
 
Have you checked the driver input when the output degrades ?
What is the value of R3 ?

MarkT

The image refers to SCT3080ALG11, which seems to be a Silicon carbide MOSFET, SCT3080AL, which
has _very_ stringent gate voltage requirements(*) - you cannot drive this without a specialist gate
driver powered at 18V.

The circuit shown will only control a logic-level silicon MOSFET.


(*) required gate drive 18V, gate abs max 22V, abs min -4V - heavy protection against
dV/dt induced voltage spikes would be needed to avoid gate breakdown as the leeway is only 4V

Compare with standard logic level MOSFET gate, usually +/-20V limit on gate voltage
and 5V nominal, so +15/-20V leeway available.  Non-logic level Si MOSFETs usually have
+/-30V limit...
[ I DO NOT respond to personal messages, I WILL delete them unread, use the forum please ]

raschemmel

Quote
Also, no mention of gate driver Power voltage (VDD . What voltage is the driver running off ?

JohnRob

MarkT is correct.  I didn't read completely through the specification the first time.

Please note the caution on the specification sheet:

Please be advised not to use SiC-MOSFETs with Vgs below 13V as doing so may cause
 thermal runaway.
Please do not PM me with thread based messages.  If your thoughts are worth responding,  the group should benefit from your insight.

raschemmel

#7
Oct 23, 2018, 12:52 am Last Edit: Oct 23, 2018, 01:11 am by raschemmel
Quote
Please be advised not to use SiC-MOSFETs with Vgs below 13V as doing so may cause
 thermal runaway.
Quote
Haven't seen any mention of chip temperature ? Have you not touched the driver chip when it starts to
degrade ?
@OP,
This is a good example of the importance of reading a datasheet thoroughly.  (something most people hate to do but skimming datasheets is counter productive).(Really, who curls up with a long datasheet for bedtime reading ? (besides Paul_B)

SCT3080AL DATASHEET



YMute

Perhaps a stronge pullup on the enable pin?  Is the decoupling ceramic and within 5mm of the package?

Oh, and is it getting hot?
Sorry, I am pretty new to this area, but what does it mean by decoupling ceramic?

And when I touched the MOSFET driver, it was not hot or getting warm.

YMute

Well ok, if we are going to troubleshoot the circuit using an oscilloscope, how about showing us the scope
screen images (before/after) by posting photos taken with your cellphone ?

Also, no mention of gate driver Power voltage (VDD . What voltage is the driver running off ?

Have you checked the driver input when the output degrades ?
What is the value of R3 ?
1.jpg is when the circuit starts, it shows some degradation already, but I think it is because I am currently using a damaged driver chip. I am waiting on new ones to come.
2.jpg is after about 30 minutes then 3.jpg happens.
And then it shows no signal.

And for driver input from Arduino Uno Rev 3, it was still abou 4.8 to 4.9 V.

And R3's value is 1.05K Ohms.

For Vdd, SCT3080AL shows 300V (?) from datasheet, but MOSFET driver, IXDN614PI, does not have Vdd value on its datasheet.



The image refers to SCT3080ALG11, which seems to be a Silicon carbide MOSFET, SCT3080AL, which
has _very_ stringent gate voltage requirements(*) - you cannot drive this without a specialist gate
driver powered at 18V.

The circuit shown will only control a logic-level silicon MOSFET.


(*) required gate drive 18V, gate abs max 22V, abs min -4V - heavy protection against
dV/dt induced voltage spikes would be needed to avoid gate breakdown as the leeway is only 4V

Compare with standard logic level MOSFET gate, usually +/-20V limit on gate voltage
and 5V nominal, so +15/-20V leeway available.  Non-logic level Si MOSFETs usually have
+/-30V limit...
Ah, I see. Thank you for pointing that out. I am not too familiar with datasheet neither... I am very new to this area.

But I still do not understand why the MOSFET driver itself does not output constant ~5V values but degrades overtime.

raschemmel

#10
Oct 23, 2018, 07:35 pm Last Edit: Oct 23, 2018, 07:43 pm by raschemmel
Why are you asking this ?:


Quote
But I still do not understand why the MOSFET driver itself does not output constant ~5V values but degrades overtime.
in view of this:

Quote
but I think it is because I am currently using a damaged driver chip. I am waiting on new ones to come.
I still do not see an answer to this question:
Quote
Also, no mention of gate driver Power voltage (VDD .

What is the supply voltage for  the driver  ?

MarkT

Sorry, I am pretty new to this area, but what does it mean by decoupling ceramic?
Is the decoupling capacitor for the driver a ceramic capacitor (it very much needs to be, and
very close to the pins of the driver).
Quote
And when I touched the MOSFET driver, it was not hot or getting warm.

Which is encouraging.

The screen shots could indicate a poor connection somewhere.  Was that the output of the driver when
loaded or not loaded?

Did you look at the supply to the driver, checking it was a good solid 5V (that driver is rated from 4.5V up).

I repeat that SiC MOSFET, if its the one you use, requires careful attention to gate driver circuitry - remember
all MOSFET gates are extremely fragile and will not tolerate over voltage for even a nanosecond, There's only
a few dozen atoms between gate and source.  It needs 18V drive without question, and it needs careful
design to limit over- or under-voltage spikes on the gate in the face of source voltage bounce due to
the self-inductance of the source lead.  A standard Si MOSFET driver is not the device for non-Si MOSFETs,
SiC nor GaN...
[ I DO NOT respond to personal messages, I WILL delete them unread, use the forum please ]

raschemmel

Just so we are all on the same page , the following comment:
Quote
The image refers to SCT3080ALG11, which seems to be a Silicon carbide MOSFET, SCT3080AL
is a reference to the MOSFET , NOT the mosfet driver, so feeling the driver is not going to tell you if
the mosfet is getting hot. Please indicate the Post # of the post where you indicate that you checked
the mosfet for overheating.


The mosfet driver is probably not working for the reasons already stated (VDD < 18v).



YMute

Is the decoupling capacitor for the driver a ceramic capacitor (it very much needs to be, and
very close to the pins of the driver).Which is encouraging.
For the 10 uF capacitor, I am using radial aluminum electrolytic capacitor, not a ceramic capacitor. And it is near the driver.


Quote
The screen shots could indicate a poor connection somewhere.  Was that the output of the driver when
loaded or not loaded?
It was loaded with Arduino's 5V TTL signal with 0.1 Hz. But the output of the driver was not connected to the SiC MOSFET.


Quote
Did you look at the supply to the driver, checking it was a good solid 5V (that driver is rated from 4.5V up).
Yes, it was almost constant 4.8~4.9V from Arduino to the driver.


Sounds like I need to find a new MOSFET and change the SiC MOSFET out.
Do you have any suggestion?
We are looking to generate pulse as high as 1 MHz.



Why are you asking this ?:
"But I still do not understand why the MOSFET driver itself does not output constant ~5V values but degrades overtime."
Even when I removed the SiC MOSFET to see if MOSFET driver would be damaged, the circuit somehow still damages or shows degraded values over time on oscilloscope.
And I could not find it or seem to understand the problem correctly.


Quote
I still do not see an answer to this question:
"What is the supply voltage for  the driver?"
For supply voltage to the driver, it is 5V from digital power output port on Arduino Uno Rev3.
It is almost like 4.9V when I probed with a multimeter.


raschemmel

#14
Oct 23, 2018, 09:23 pm Last Edit: Oct 23, 2018, 09:33 pm by raschemmel
Let's review:

This comment:
Quote
Is the decoupling capacitor for the driver a ceramic capacitor (it very much needs to be, and
very close to the pins of the driver).
Is a reference (obviously) to a decoupling capacitor (typically 0.1uF , ceramic, but could be mylar),
NOT a reference to a FILTER capacitor (typically 10uF to 100uF)

The following comment:

Quote
For the 10 uF capacitor, I am using radial aluminum electrolytic capacitor, not a ceramic capacitor. And it is near the driver.
Should be:
Quote
I don't have a decoupling capacitor. All I have is a filter capacitor. I will get a 0.1uf/ceramic or mylar decoupling capacitor and install it as close to the mosfet driver as possible.
Capactive Reactance XC = 1/(2*Pi*f*C)

Case 1:
------------------------------------
f = 0.1 Hz     (100mHz from scope screenshot)


SCOPE SCREENSHOT:

Pi = 3.14159265359
C = 10E-6 uF (10uF from OP's circuit)

Let C = 10uF (10E-6) F

XC= 159.1 kohms (CURRENT CASE AS GIVEN FROM SCHEMATIC
                                                                       AND SCOPE SCREENSHOT)
-----------------------------
Case 2:
------------------------------------
f = 0.1 Hz     (100mHz from scope screenshot)
Pi = 3.14159265359
Let C =  0.1E-6 uF (0.1uF (standard decoupling cap value)
XC= 15.1 Mohms
-----------------------------
Case 2:
------------------------------------
f = 1 Mhz  (1 Mhz frequency proposed by OP in post)
Pi = 3.14159265359
Let C =  0.1E-6 uF (0.1uF (standard decoupling cap value)
XC= 1.59 ohms (CASE FOR 0.1uF Cap at 1 Mhz switching frequency)
-----------------------------


CONCLUSION:
If you stop using 0.1 Hz and start using a higher , more realistic frequency (like 100 kHz)
AND add a 0.1 uF DECOUPLING CAP (in addition to current 10uF FILTER CAP), the decoupling
cap will pass high frequency spikes to ground , reducing noise.

Go Up