Maker Pro
Maker Pro

BJT's Moisture Sensitive Devices storage?

H

Hammy

Jan 1, 1970
0
I just bought some BJT's OnSemi's low Vcesat high gain NJT4031NT1G and
complementary NJT4030PT1G.

NJT4031NT1G Data sheet

http://www.onsemi.com/pub/Collateral/NJT4031N.PDF

NJT4030PT1G Data sheet

http://www.onsemi.com/pub/Collateral/NJT4030P-D.PDF

They came with an advisement "Caution this bag contains Moisture
Sensitive Devices" Level #3

Where level 3 is

"floor life out of bag"
168 hours <30 C/60% RH

Does this mean that these BJT's would only last 168 hours shelf life
when taken out of the bag? This doesn't make any sense what if you use
these in a circuit that only see's intermittent use will the BJT specs
deteriorate?

I have no immediate need for these I just got them to test them out
switching speed,vcesat and gain.

I usually store my smd components rewraped in antistatic bags and
labeled in coin envolopes. I dont revacum seal them though. Would this
method still work for components with Moisture warnings?
 
R

Rich Webb

Jan 1, 1970
0
I just bought some BJT's OnSemi's low Vcesat high gain NJT4031NT1G and
complementary NJT4030PT1G.

NJT4031NT1G Data sheet

http://www.onsemi.com/pub/Collateral/NJT4031N.PDF

NJT4030PT1G Data sheet

http://www.onsemi.com/pub/Collateral/NJT4030P-D.PDF

They came with an advisement "Caution this bag contains Moisture
Sensitive Devices" Level #3

Where level 3 is

"floor life out of bag"
168 hours <30 C/60% RH

Does this mean that these BJT's would only last 168 hours shelf life
when taken out of the bag? This doesn't make any sense what if you use
these in a circuit that only see's intermittent use will the BJT specs
deteriorate?

I have no immediate need for these I just got them to test them out
switching speed,vcesat and gain.

I usually store my smd components rewraped in antistatic bags and
labeled in coin envolopes. I dont revacum seal them though. Would this
method still work for components with Moisture warnings?

That limit is "time to reflow." If you do reflow them, you should bake
them at the recommended temperature profile to drive out moisture before
subjecting them to reflow temps. Good dope at:
http://www.intel.com/design/packtech/ch_08.pdf
 
H

Hammy

Jan 1, 1970
0
Theoretically, a surface-mount device can absorb moisture and crack
when reflow soldered, so they want them kepy dry until then. I've
never seen it happen.

They'll last about forever once on a board.

John

Thanks I didn't want to ruin them; I've also gotten some other
components with similar warnings. I guess since I solder by hand it's
not a problem.

Did you see the specs on these fig. 6 (NJT4031NT1G Data sheet) Vce sat
of less then 300mV for 3A of collector current with only 25mA base
drive. Perfect for direct drive off a uC or a cheap 3A push-pull Fet
driver. I know zetex makes good ones also but they are pretty pricey
these only cost 0.30 cents each cut tape. Why so cheap what am I
missing?
 
P

Phil Allison

Jan 1, 1970
0
"Hammy"
Did you see the specs on these fig. 6 (NJT4031NT1G Data sheet) Vce sat
of less then 300mV for 3A of collector current with only 25mA base
drive. Perfect for direct drive off a uC or a cheap 3A push-pull Fet
driver. I know zetex makes good ones also but they are pretty pricey
these only cost 0.30 cents each cut tape. Why so cheap what am I
missing?


** A couple of things - you have equated the "typical" performance data
curves with guaranteed minimis and ignored the heat dissipation problems of
such a small package.

A worst case (ie low Hfe) sample might well take several times your 25mA
figure for the same Vce at 3 amps PLUS the dissipation in any case will be
almost 1 watt.

Read carefully what the data has to say about this and note that even with a
1 sq inch of copper area acting as a heatsink, the temp rise above the local
ambient of the chip is 64 degrees C for each watt.



...... Phil
 
H

Hammy

Jan 1, 1970
0
** A couple of things - you have equated the "typical" performance data
curves with guaranteed minimis and ignored the heat dissipation problems of
such a small package.

A worst case (ie low Hfe) sample might well take several times your 25mA
figure for the same Vce at 3 amps PLUS the dissipation in any case will be
almost 1 watt.

Read carefully what the data has to say about this and note that even with a
1 sq inch of copper area acting as a heatsink, the temp rise above the local
ambient of the chip is 64 degrees C for each watt.



..... Phil
Figures it was too good to be true.

I did hook one up driving it off a 556 timer at 144kHz 58% duty and it
is reasonably fast at turn-on but the turn off leaves a bit to be
desired it takes about 1.4uS to turn-off!

For 1.2A collector pulses it was giving 0.08V collector emitter with
about 10mA at the base driven from the 556.
 
J

Jasen Betts

Jan 1, 1970
0
I did hook one up driving it off a 556 timer at 144kHz 58% duty and it
is reasonably fast at turn-on but the turn off leaves a bit to be
desired it takes about 1.4uS to turn-off!

For 1.2A collector pulses it was giving 0.08V collector emitter with
about 10mA at the base driven from the 556.

bypass some of the base resistor with a capacitor.
that'll suck charge off the base causing faster turn-off.
 
H

Hammy

Jan 1, 1970
0
Yes there is.

I just took one out and used my vacum sealer to reseal them,leaving
the desiccant in there.

I also got some NCP1203 Controllers they also came with a similiar
warning. But the FMMT618 and FMMT617 didnt come with the mousture
warning.

Is this a concern with all SMD or just some ?

I handsolder so I dont think I really have to worry about heating one
enough to crack it or should I ?
 
J

Jamie

Jan 1, 1970
0
Jasen said:
bypass some of the base resistor with a capacitor.
that'll suck charge off the base causing faster turn-off.
I find the Vce(Sat) figure given (0.08) hard to grasp for
that component. Maybe some capacitance/Inductance in the transient
cycle causing that reading to be low. I bet when the BJ was turned off
there was a small arch Peak on the high side.

http://webpages.charter.net/jamie_5"
 
H

Hammy

Jan 1, 1970
0
bypass some of the base resistor with a capacitor.
that'll suck charge off the base causing faster turn-off.

Yes I've been doing some reading thats what I'll try next.A Baker
clamp defeats the purpose of a low Vcesat BJT but I've found a few
things to try.

If I can get good transtions to about 100kHz I'll be happy,thats the
typical maximum switching speed I usually wind up with.

They are reasonable priced so even if I have to spend another 5 or 10
cents on a driver cct it's worth it.Even at the worst case Vce
sat(0.3V) and hfe (100) for 3A.
 
H

Hammy

Jan 1, 1970
0
I find the Vce(Sat) figure given (0.08) hard to grasp for
that component.

Your probably right I just read the measurment my scope spat out. I
get more accuarate readings when zooming in but 0.1V is probably more
likely given the max they spec 0.150V. I'll find out for sure when I
do it again.
 
H

Hammy

Jan 1, 1970
0
I find the Vce(Sat) figure given (0.08) hard to grasp for
that component. Maybe some capacitance/Inductance in the transient
cycle causing that reading to be low. I bet when the BJ was turned off
there was a small arch Peak on the high side.

http://webpages.charter.net/jamie_5"

More accurate measurements;

Vbe = 0.92V

556 Vout = 8.9V (The timer is supplied with 12V.)

I have a 1.2k 1% metal film resistor between the timer and the base.
So the base is getting 6.67mA.

Zooming right up gives a Vcesat of 182mV. Still pretty respectable for
only 6.67mA of base current.

To answer your next question yes there is a one volt pp ripple on the
12V rail but I'm using a 470uF polymer cap right by the load the over
and undershoot are essentially brief spikes so the collector current
is 1.2A. The load is two 5R 5W resistors 5% so maybe worst case 1.1A.
 
J

Jamie

Jan 1, 1970
0
Hammy said:
More accurate measurements;

Vbe = 0.92V

556 Vout = 8.9V (The timer is supplied with 12V.)

I have a 1.2k 1% metal film resistor between the timer and the base.
So the base is getting 6.67mA.

Zooming right up gives a Vcesat of 182mV. Still pretty respectable for
only 6.67mA of base current.

To answer your next question yes there is a one volt pp ripple on the
12V rail but I'm using a 470uF polymer cap right by the load the over
and undershoot are essentially brief spikes so the collector current
is 1.2A. The load is two 5R 5W resistors 5% so maybe worst case 1.1A.

That's more realistic :)

One thing you may want to take into consideration is, BTJ's tend to
loss their Beta/Hfe gain when operating in cold environments.

To test this, I've sprayed BTJs with component cooler and then take a
few readings. This could lead you into changing the bias resistor to a
slightly lower value to accommodate for this.

Happy New Year..

http://webpages.charter.net/jamie_5"
 
E

Eeyore

Jan 1, 1970
0
Hammy said:
Figures it was too good to be true.

Practical example. A Chinese company did a partial ODM designed product for us. I
refined it. They'd used some SOT 23 devices, MMBT3904 and complement IIRC as FET
gate drivers. We had two explode on test.

In went the 2N3904 and its complement.

Graham
 
Top