Early Voltage problems

Tue Aug 24, 2010 3:13 pm   

I'm trying to model a 6n139 split-darlington opto-coupler in Spice. I'v
collected hFE data on the Q1 (~45) and Q2 (~100) of my tube of 25 chips
but I've run into a problem with the Early voltage of Q2.

I set base currents at 5~~100 uA using an LM334 and 15 resistor chain, use
a 3V~~12V switched PSU for Vcc.

When you plot iCollector against Vcc for different values of iBase, th
tangents are supposed to converge, albeit waaay off to the left.

Well, mine don't. At 10 uA (μA), Early voltage is ~200. At 20 uA ~160, 5
uA ~100, 70 uA ~85 and 100 uA ~65. I've tried ignoring the 3 & 4.5
data-points, but the tangents still don't converge...

Any ideas where I've gone wrong ??
(I'm a hobbyist, not a student.)

---------------------------------------
Posted through http://www.Electronics-Related.com

Thu Aug 26, 2010 7:12 pm   

On Tue, 24 Aug 2010 09:13:13 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


Post a schematic of, or a link to, your fixture.

You're probably getting some saturation or don't fully have the driver
transistor off.

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Spice is like a sports car...
Performance only as good as the person behind the wheel.

Thu Aug 26, 2010 8:40 pm   

On Aug 24, 7:13 am, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:

Nice effort. Not too many explore this; and you may be finding part
of the reason.

You say that you set the "base current" - but which? The output
transistor alone?

If it is just the output transistor (else all bets are off)- the
process used may be optimized for its optical response characteristics
- and have really soft saturation characteristics (which you see as a
current-dependent Early voltage). Try running the same tests on a
known good transistor (say, 2N2222) to see if there's something wrong
with your test setup, calculations, or if it is the part.

HTH--

Sun Aug 29, 2010 2:25 am   

I got non-convergent tangents for Q2, giving VAF ranging from 70~~170
After much hair-tearing and deep searching, I've learned that this is
'well-known' issue in devices with small areas and/or self-heating.

In these conditions, the Early voltage is current dependent, and does no
conform to Gummel-Poon model.
;-(
So, I'll probably leave Q2 VAF at the default 100. For Q1, I'll use VAF
17 (seventeen) from Avago's published 6n136 model...

---------------------------------------
Posted through http://www.Electronics-Related.com

Sun Aug 29, 2010 3:28 am   

On Sat, 28 Aug 2010 20:25:38 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


Does not compute. For a single chip Darlington, VAF of BOTH devices
should be same. I question validity of characterization.

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Spice is like a sports car...
Performance only as good as the person behind the wheel.

Sun Aug 29, 2010 2:46 pm   

"I question validity of characterization."

Jim, I was surprised, too !!

In the absence of a manufacturer's model, I was able to directly determin
hFE for Q2 as it had accessible EBC. My Peak Atlas (PA) mini componen
tester gave 25 results with an average of 101.1 and an SD(n-1) of 5.0

I'm fairly confident of the LED==>photo-diode CTR being ~0.0015 (0.15%) a
this was quoted in H-P Opto-electronics Applications Manual (c)1977, p3.7
as valid for a range of devices and it matches the ~0.16% of relate
6n136's data-sheet and spice model.

Connecting LED cathode to Q2 emitter and commoning Q1~Q2 collectors let th
PA recognise entire device as a Darlington. That gave an overall mean hF
of 7.3

Given those, a Q1 hFE of ~45 just falls out of the mix...

D'uh, like the Early Voltage being non-convergent for small-area device
--Due to 'current crowding' ??-- is it improbable that 'buffer' Q1 i
optimised for speed and sensitivity, while Q2 carries the bulk of the gai
?

Unfortunately, Avago (was Agilent, was H-P) aren't talking...

FWIW, I did Q2's 'Early' curves on one chip at 5,6,7,8,9,10, 20, 30, 40
50, 60, 70, 80, 90 & 100uA iBase and 3, 4.5, 6, 7.5, 9 & 12 VCC. The 100u
curve had both a marked 'soft shoulder' and a hint of approachin
breakdown. With or without dubious end-points, the tangents did no
converge. At least I've found a likely basis for that anomaly...

---------------------------------------
Posted through http://www.Electronics-Related.com

Sun Aug 29, 2010 4:28 pm   

On Sun, 29 Aug 2010 08:46:22 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


If you're seeing current crowding, parameter "IKF", then get Early
effect "VAF" from low currents, make model, then go back and determine
IKF. If your simulator can handle avalanche parameters, add those as
well.

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Spice is like a sports car...
Performance only as good as the person behind the wheel.

Sun Aug 29, 2010 10:04 pm   

Thank you.
Nik

---------------------------------------
Posted through http://www.Electronics-Related.com

Sun Aug 29, 2010 10:20 pm   

On Sun, 29 Aug 2010 16:04:11 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


I should have also pointed out that BF will be the same for both
devices, but each device will have its own IKF, IS and NE, and other
area (and geometry) dependent terms.

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Spice is like a sports car...
Performance only as good as the person behind the wheel.

Mon Aug 30, 2010 2:22 pm   

FWIW, here are the resistor values I used to set 5~~100 uA iBase.

My supplier only stocked low-value metal-film 1% resistors in E12 serie
and an LM 334 has a nominal +/- 3% error, so I measured the delivere
current prior to connecting the opto-coupler.

LM334 mA = 67.7 / R

TargetmA R_step R_Cum I_nominal Error % Actual uA
0.100 680.0 680.0 0.100 -0.4 99.9
0.090 82.0 762.0 0.089 -1.3 89.2
0.080 91.0 853.0 0.079 -0.8 79.6
0.070 110.0 963.0 0.070 +0.4 70.5
0.060 160.0 1123.0 0.060 +0.5 60.4
0.050 240.0 1363.0 0.050 -0.7 49.8
0.040 330.0 1693.0 0.040 -0.0 40.1
0.030 560.0 2253.0 0.030 +0.2 30.1
0.020 1100.0 3353.0 0.020 +1.0 20.2
0.010 3300.0 6653.0 0.010 +1.8 10.2
0.009 910.0 7563.0 0.009 -0.5 9.0
0.008 910.0 8473.0 0.008 -0.1 8.0
0.007 1200.0 9673.0 0.007 -0.0 7.0
0.006 1600.0 11273.0 0.006 +0.1 6.0
0.005 2200.0 13473.0 0.005 +0.5 5.0


---------------------------------------
Posted through http://www.Electronics-Related.com

Mon Aug 30, 2010 6:54 pm   

Jim, I seem to have goofed with that BF split. The photo-diode, Q1 & Q2 ar
surely monolithic hence have similar junctions in all but size. FWIW,
comment in Zetex model info (see below) was only reference Google coul
find for Darlington gain splits...

Yes, I know that transistors on-chip may be very well matched, it is wha
makes long-tail pairs and current mirrors so predictable, but how do TO22
Darlingtons contrive dissimilar Driver and Power stages ??

Okay, I've a Plan_B for driving my 6n139s' LED with known current fro
LM334 and measuring Q1's collector current. Take a few days to source th
dropper chain parts...

===(c)Zetex===
Darlingtons
These are subcircuits using a standard transistor model. A Darlington mode

is shown as follows:
*
*Zetex FZT605 Spice Model v1.0 Last revision 27/04/05
*
.SUBCKT FZT605 1 2 3
* C B E
Q1 1 2 4 SUB605
Q2 1 4 3 SUB605 3.46
*
.MODEL SUB605 NPN IS=4.8E-14 BF=170 etc.
.ENDS FZT605
*
*$
Note:
Because Zetex Darlingtons are monolithic, the two transistors used ar
identical in all respects other than size. (The number at the end of the Q
line multiplies the size of the SUB605 transistor by 3.46 - the ratio o
the areas of the input and output transistors for this device).


---------------------------------------
Posted through http://www.Electronics-Related.com

Mon Aug 30, 2010 7:15 pm   

On Mon, 30 Aug 2010 12:54:16 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


I knew that :-)

What issues are you having? Matching Input (LED) current to output
curves? Perhaps both driver and output device are exposed to the
light?? Seems odd, but that could really !@#$% up your results.

Way back when I did optical chips, we masked off everything but the
desired opto-active device with an extra layer of metal.

Modern processes actually have an opto-blocking layer in the process,
and layers that enhance specific wavelengths (see XFAB for examples).

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Spice is like a sports car...
Performance only as good as the person behind the wheel.

Fri Sep 03, 2010 8:21 pm   

The basic problem with modelling the 6n139 is that Q1's base is no
accessible, so *only* the product LED==> photo-diode CTR x Q1 hFE i
available...

(CTR = Current Transfer Ratio ;-)

Or so I thought...

Last night, I had a rush of blood to the brain: What if I fed the LED a fe
milliamps and put my DVM across Q1's emitter and collector ? No Vcc, jus
the photo-diode's output ? Would its photo-current flow through Q1'
base/emitter PN ?

So, today, I tried it. I bridged Q1 emitter to Q2 emitter as latter i
connected to substrate. I put an LED in series with the opto's LED to se
current was flowing. A 9V battery and 680R, 1K0 or 1K8 fed 9.05, 6.21 o
3.49 mA by DVM. I measured 53.4, 35.4 and 18.6 uA photo-current across Q
emitter and collector.

These gave CTR of 0.0059, 0.0057 & 0.0053, equivalent to ~0.5 %.

This is approx three (3) times the CTR from the old H-P reference, thric
the quoted figure for the data-sheet and model of 6n136...

If this ratio is representative, it means that a Q1's hFE is only ~15, i
marked contrast to the Q2s' measured hFE of ~100...

D'uh, can you see any hole in my methodology or logic ??

---------------------------------------
Posted through http://www.Electronics-Related.com

Fri Sep 03, 2010 9:41 pm   

On Fri, 03 Sep 2010 14:21:56 -0500, "Nik"
<Nik_2213_at_n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:


Yes.

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Democrats are Living Proof that Government IS the Problem!

Sat Sep 04, 2010 1:44 pm   

Base-Collector junction. Why so much went via Base-Emitter is probably dow
to the 'small' geometry...

D'uh, my bad...

---------------------------------------
Posted through http://www.Electronics-Related.com