Buyer Beware

[Pooh Bear]

Yes... the beta might be a lot higher, for instance.

Speaking of which, are the TL071 and TL081 op-amp now the same chip?
Originally they weren't, but nowadays I can't find any difference between
them.

I've noticed that the 081 now sports the same noise figure as the 071.

I guess TI decided it wasn't worth running 2 separate lines that were so
similar. The 071 used to be pricier too.

Graham

 

[Rich Grise]

Right. But the person who designed your stereo in 1964 may not have been
taught that way...

When I worked at Radio Shack in the repair department in 1972[1], we
always replaced blown output transistors with 2N3055s. ("Oops - I forgot
to discharge the scope probe before I probed the base [THUMP!][2] - Do we
have any XXXXX in stock?" "Ah, just throw in a couple 3055's.")

Cheers!
Rich

[1] Yes, Radio Shack actually had a repair department back in those
days, with marginally competent techs. When I got there, they had a
3-month backlog. In about a month, we were caught up. When the boss
threatened to put me in charge while he went on vacation, I quit and
re-enlisted in the Air Force.
[2] Yes, a 10:1 scope probe is a capacitor, and it can get charged,
when you're scoping a 90V rail. Transistors don't like having that
cap discharged through their base.

 

[Watson A.Name - \Watt Sun, the Dark Remover\]

We have been more fortunate with ebay, not yet had a bad one other than a
displaced galvanometer movement probably caused by postal handling.

It does depend a lot on what you are buying and how much you have paid IMO. The
guy who gets a reasonable price for his sale is more likely to pack properly etc
than the guy who expected $15 and got $2 for example.

I had one like that last week or so, he wanted $15 for posting some packs of
plotters pens internally in the USA, while we were able to point out to him that
$5.50 was more than sufficient.

Takes all sorts! )

Peter

Speaking of which...
Well, I _finally_ got Tucker (tucker.com) to agree that they weren't
allowed to charge tax on shipping. It took several emails and phone
calls, and finally I got hold of the lady in accounting, who said she'd
call me back, and finally she did, and allowed me to not pay taxes on
shipping. I could've showed them the Calif Board of Equalization FAQ
that says that shipping (the FAQ calls it transportation) is not
taxable. And I could've showed them the federal law that says it's
illegal, for example, for a seller in Taxes to apply Texas tax
regulations to a buyer who's a resident of California.

I hope they get their act together and don't subject any more buyers to
those hassles.

 

[Watson A.Name - \Watt Sun, the Dark Remover\]

2N4264 has an order of magnitude less switching storage time than the
2N3904 and one fifth the current fall time. It doesn't suffer from
gain reduction or desaturation above 100mA - being characterised to
200mA with low VCEsat. The trade-off is operating voltage, being
useful only in <12V systems.

The 2N4124 is not really characterized for switching purposes.

2N4264 isn't an 'audio' device, but has reasonable small-signal
characteristics for use in amplification.

What is your motive in spreading mis-information? The seller's motive
is obvious - he's selling.

I sent the seller an email _again_ this morning complaining that another
auction was wrong. The first one had the specs _and_ the banner for the
MPS65<something> in the auction for the 2N4264, which was deceptive.
The auction I found this morning had the MPS banner replaced with
2N4264, but with the specs for the MPS still there, which is even more
deceptive.

 

[Watson A.Name - \Watt Sun, the Dark Remover\]

Right. But the person who designed your stereo in 1964 may not have been
taught that way...

In 1964 the common audio amp transistors were 2N3391A, 2N3900, and a few
others from Fairchild, and the 2N5088 and similar. They were cheap, and
did the job. The opamps used in audio amps that I can remember were
specialty chips made just for audio, often with oddball pinouts and/or
in a special package. I guess that opamps were not low noise back then.

 

[Phil Allison]

"mc"
Phil Allison:

Yes... the beta might be a lot higher, for instance.

** In the case of the MJ802/4502 pair, their beta remained high at low
collector currents (ie under 100 mA), unlike the MJ 15003/4 substitutes.
Also, the Vbe/ Ic characteristic is dissimilar, causing large imbalance in
Ic when devices are operated in parallel groups. Also, the Ft is greater
which can lead to parasitic oscillations.

It is to avoid the above kinds of issues that folk want to buy the original
types and pay the extra - then find they have been duded.



............. Phil

 

[Phil Allison]

"Phil Allison"

- then find they have been duded.

** That should be "dudded" - an Aussie colloquialism for " sold a pup ".



............ Phil

 

[Graham W]

I thought it had something to with commercial vs. industrial
applications, possibly temp. range, but to be honest, that's a WAG.

Nah - each type was (in my '82 Bifet Design Manual spec sheets)
available in M, I, &C suffixes relating to mil, ind, & comm grades.
The 07x was 18 nV per root.Hertz while the 08x was 25.

 

[Don Klipstein]

In "Seat Of The Pants Design School" they taught me to design such that
beta is inconsequential anyway, as long as it's "enough".


I thought it had something to with commercial vs. industrial applications,
possibly temp. range, but to be honest, that's a WAG.

What I heard back in the 80's was that TLO71 was supposed to have lower
noise than TLO81. The difference was mainly in lower impedance circuits,
since the noise rating for TLO81 I calculated to equal the thermal noise
of roughly a 15K resistor.
I did notice a difference in stability in some circuits feeding
capacitive loads when wired for lower gain, such as shielded cables - the
TLO81 was better. (I have found things are better if you add a resistor
in series with the conductor of the shielded cable being fed by an op amp
wired for low or unity gain.)

(Actually my experience was with TLO84 and TLO74.)

I wonder if nowadays they make ones that meet the specifications of
both?

- Don Klipstein ([email protected])

 

[Don Klipstein]

Thanks for that... it makes sense to indicate an item as "repairable" when
they know that it doesn't work, but it would make more sense to indicate an
item as "unchecked" if that is the case and its condition is unknown. Oh
well.. guess they are too set in their ways to change...

Makes me think "repairable" is like a bill being "payable". The item
likely needs repairs. There is no statement as to how easy this will be.
If they took the time trouble to find out what went wrong, they probably
would have taken the probebly less time to actually fix it.

If I was famaliar with the item being sold as "repairable" and know of
common and easy to fix failure modes that probably necessitated repairs
and that the seller probably does not know about or that the seller
probably would not bother to fix, then I would buy it.

- Don Klipstein ([email protected])

 

[Phil Allison]

"Don Klipstein"

What I heard back in the 80's was that TLO71 was supposed to have lower
noise than TLO81. The difference was mainly in lower impedance circuits,
since the noise rating for TLO81 I calculated to equal the thermal noise
of roughly a 15K resistor.

** The original noise spec for the TL081/2/4 was given as 47 nV per rt Hz.

This equates to a 135 kohm resistor.

( R = N*2 / 4.K.T )



............ Phil

 

[Pooh Bear]

"Don Klipstein"

** The original noise spec for the TL081/2/4 was given as 47 nV per rt Hz.

This equates to a 135 kohm resistor.

( R = N*2 / 4.K.T )

Rubbish - the original noise spec for the 081/2/4 was 25nV/rtHz.

18nV for the 071/2/4

and 42nV for the 061/2/4.

I have the original data sheets in case you feel like arguing the point !

Graham

 

[Joel Kolstad]

Rich,

In "Seat Of The Pants Design School" they taught me to design such that
beta is inconsequential anyway, as long as it's "enough".

Hey, in digital school the kids are taught to use a part that's "fast enough,"
but there are many cases reported where a digital design fails because -- over
time -- the parts get faster (i.e., they're die shrunk but still labeled the
same) and suddenly the faster edges start causing signal integrity problems!

I imagine one can site similar cases in the analog world. I.e., a transistor
oscillator with a gain an order of magnitude higher than what was originally
designed for I could see as ending up with significant purity problems,
especially if the design was cost optimized so let the oscillator self-limit
rather than forcibly designing it to do so.

---Joel

 

[Phil Allison]

"Pooh Bear"

Rubbish - the original noise spec for the 081/2/4 was 25nV/rtHz.

18nV for the 071/2/4

and 42nV for the 061/2/4.

I have the original data sheets in case you feel like arguing the point !

** I have in front of me the brochure published by Texas Instruments on the
release of their BiFet op-amp series entitled :

" Three BIFET OP AMP Series from Texas Instruments - Bulletin CL-309 "

It is also marked: " Copyright 1977 Texas Instruments Incorporated " -
same year the devices were released.

In the "Specification Comparison " tables AND in the detailed data for the
TL081 series, the figure given is 47 nV.





.......... Phil

 

[mc]

Rich,


Hey, in digital school the kids are taught to use a part that's "fast
enough,"
but there are many cases reported where a digital design fails because --
over
time -- the parts get faster (i.e., they're die shrunk but still labeled
the
same) and suddenly the faster edges start causing signal integrity
problems!

I imagine one can site similar cases in the analog world. I.e., a
transistor
oscillator with a gain an order of magnitude higher than what was
originally
designed for I could see as ending up with significant purity problems,
especially if the design was cost optimized so let the oscillator
self-limit
rather than forcibly designing it to do so.

Yes... We ham radio people are cautioned not to build RF power amplifiers
with transistors rated for a frequency much higher than we're actually
using, because they're apt to oscillate.

 

[Pooh Bear]

"Pooh Bear"


** I have in front of me the brochure published by Texas Instruments on the
release of their BiFet op-amp series entitled :

" Three BIFET OP AMP Series from Texas Instruments - Bulletin CL-309 "

It is also marked: " Copyright 1977 Texas Instruments Incorporated " -
same year the devices were released.

In the "Specification Comparison " tables AND in the detailed data for the
TL081 series, the figure given is 47 nV.

.......... Phil

They changed it damn fast in that case ! I'm quoting from Bulletin DL-S 12484
Feb 1977 - admittedly revised June 1978.

It's now 18nV btw.

Graham