Transistor tester

I'm just finishing up an interesting job for a group in the Department of Chemistry at Notre Dame. It's a laser noise canceller that works down in the 10 nA-50 uA range, about 100 times lower current than the original model.

Noise canceller performance depends pretty well completely on BJT diff pairs and cascodes performing as Ebers and Moll predict, so that the fluctuations see exactly the same gains as the DC.

The two main sources of error are log nonconformance, which makes the diff pair's large- and small-signal current division ratios different, and beta error (1/h_FE - 1/h_fe), which causes the same problem in the base/collector split of the emitter current.

I built a tester that works over an emitter current range of 100 pA to

100 mA, and seems to do a good job of measuring those two parameters directly. (You have to be careful about junction temperature in the log conformance measurement, of course.)

The vellum got a little bit munched in the process, but the scan is still pretty readable and has some points of interest. (It even has a 555.)

formatting link

Building it would have taken half as much time if I'd used a bigger box!

Early results indicate that modern RF transistors are _much_ better than old ones at low I_C. Both BFT25As and the Infineon HFA3xxx parts have betas of well over 20 at 100 pA I_C.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs
Loading thread data ...

Good low current beta performance is directly related to clean processing and surfaces. So one would hope that "modern" is better ;-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

I'm going to try with some SiGe:C ones on Monday, just for grins. I bet I can get to at least 100 nA I_E before they start oscillating. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

I've done some chip designs with mixed CMOS and SiGe with no particular problems, but I'd guess one loose in your hands will be one squirrelly creature ;-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

I use them as cascodes for pHEMTs, for which they work brilliantly, as long as you put the right little ferrite bead in the right place. ;)

The combination of really low noise, high transconductance, very high beta, and effectively infinite Early voltage completely fixes the crappy soft drain characteristic of a pHEMT.

(That was for a built-up amplifier to make shot-noise limited measurements of 1-nA currents in a bandwidth of 100 MHz. [The shot noise limit is N/(2B) per hertz, where N is the mean electron arrival rate--in this case 31 electrons.] It almost made it--it comes in about

4 dB off the pace, but that's at least as good as any built-up circuit I've ever heard of. It needed a combination of 0.3 nV/sqrt[Hz] noise out past 100 MHz, and an input capacitance below 1 pF.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

My use was in 5.5GHz down-converters. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

nt

e

ff

e

t,

he

My impression was that the major source of log non-conformance was base res istance. Emitter resistance creates the same sort of problem but is usually smaller.

to

log

a

Two transistor astables are cheaper ...

It displays side-ways on, and Firefox didn't want to print it out. Where di d the transistors plug in?

box!

Always does.

than

ve

formatting link

makes the BFT25 a rather low current transistor, not to mention low power - 32mW isn't much. Farnell stocks them for $A0.44 in small quantities

The HFA part numbers used to Be Harris and are now Intersil. The relevant p art numbers might be HFA-3101, -3102, -3127, -3128, -3134 and and probably grew out of the -3046 and -3096 that started off with RCA.

Which parts do you fancy, and why? I always liked the HFA3101 Gilbert Cell, but never found an excuse to use one.

Elsewhere you have posted "I use them as cascodes for pHEMTs, for which the y work brilliantly, as long as you put the right little ferrite bead in the right place. ;) "

My experience with GHz wide-band transistors was that I couldn't stop them oscillating except by sticking about 33R in series with the base connection , right up against the base tag, but I can see how a well-chosen ferrite be ad would do the same job, rather more expensively.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

It's not too often you see a 100 G resistor called out. :) Are the switches that select the resistors near the (DUT) base and emitter connections rotary switches, or jumper pins, or...? Seems like you might have to use a gold-plated or otherwise fancy switch at some of those teeny currents.

Why are the supplies asymmetrical? (+8.6 V, -7.5 V) In most of the analog-y op-amp-y stuff I see that has both positive and negative supplies, the supplies are symmetrical.

Matt Roberds

Reply to
mroberds

In Foxit, "View/Rotate View/Clockwise".

Reply to
John S

On a sunny day (Sat, 06 Jul 2013 21:46:15 -0400) it happened Phil Hobbs wrote in :

Nice diagram, I look much longer at the hand written ones than at the computah generated ones, it conveys some of the spirit of the author.

Reply to
Jan Panteltje

Yes, got any more you can share Phil?

Need some holiday reading material, and they don't seem to be making any more Jim Williams app notes... :(

--

John Devereux
Reply to
John Devereux

On 07/07/13 11.50, John Devereux wrote: ...

Hi John

There are plenty of nostalgic stuff here:

formatting link
formatting link

E.g.:

formatting link

formatting link
formatting link
formatting link
formatting link
formatting link

/Glenn

Reply to
Glenn

Thanks, good stuff but now a little bit *too* old school for me! :)

--

John Devereux
Reply to
John Devereux

e

No Foxit.

Downloading the .pdf file and opening it with Acrobat gave me an image that I could print. Diabolical contrast, and I'm still at loss where Phil might plug in his transistors. The elements are recognisable enough - I've put t ogether logging and anti-logging circuits (though not recently) - but it's not the most intelligible circuit diagram I've ever seen.

It tends to confirm my negative feelings about physicists and electronics. It's not a particularly horrible example - I been more peeved by neater cir cuit diagrams published in Rev. Sci. Instruments

Sloman A.W. ?Comment on ?Modular digital box-car for applications in pu lsed laser spectroscopy? Review of Scientific Instruments, 67 3763-4 (19

96)

but it's not a great advertisement for the profession.

--
Bill Sloman, sometime physical chemist,  Sydney
Reply to
Bill Sloman

I misread Firefox as Foxit. My apologies.

Reply to
John S

Nice? What's a LM2954? I assume a simple-switcher device with maybe a typo in it. My suggestion would be to add a 0.1uF ceramic in parallel to the 330uF and the two 270uF caps.

Plus maybe a wee ferrite bead and a Murata EMC suppressor where the base connects, since you are only interested in low frequencies and don't want the thing to oscillate (they can do that "silently").

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

I didn't spend much time on the schematics. Only thing I found funny was to use a switcher in a transistor tester, a linear supply would be less noisy

Regards

Klaus

Reply to
Klaus Kragelund

Depends on the beta. Most devices that I've measured have high enough beta that R_EE' dominates. That's certainly true for the HFA parts, whose R_EE' is a fairly horrible 0.8 ohms or so.

Thought you'd appreciate that. ;)

It's a B-size scan. Most PDF readers let you rotate the view--I use Foxit, which works fine. The transistor attaches to the flags marked "E", "B", and "C".

This job needs both polarities, so I'm going to use the HFA3096. The diff pair ones are nice because they reduce the emitter inductance.

It's a pity the package parasitics are so horrible--I think it was Fred Bartoli who pointed out that the bare die has half the interelectrode capacitance of the packaged part. A SOT-16 is not the package you want for this job.

Beads are cheap, especially compared with a $3 transistor array, and they don't affect the low-frequency accuracy. A 33 ohm resistor in the base would hurt the log conformance, which I care about a lot. (This particular device only has to go up to about 100 uA, but that's still enough for a few tenths of an ohm of effective emitter resistance to be a fairly serious performance limiter.) A good noise canceller can get up to 70 dB RIN suppression at low frequency, which requires the phase shift between the two currents to be less than 0.0003 radians and the log conformance to be in the same range.

Getting that sort of log conformance at higher photocurrents requires tweaking out the extrinsic emitter resistance R_EE' by applying a small amount of positive feedback to the base.

You can get 50 dB or a bit more using separate transistors and no tweaks, which is usually better than good enough.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

I'm not too worried. Your negative feelings about physicists get lost in your negative feelings about everything else. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The old hand-inked, color Tek schematics were beautiful. Some had tekdoodles.

Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.