Rant - Bad (Chinese) engineering

Maybe in some companies. In others, they might rent or farm the work out.

In the mid-'70s I bought a Tektronix signal processing system, that was really not much more than a PC and digital scope, today, but was well over $150K, at the time. I bought a second system a few years later.

I saw a Tek seminar recently on their 9000 class scopes. 30-40GHz, tons of triggering options, 300k$. A 40GHz probe costs 20k$. I surely am not going to have one, but someone's buying them...

Lets see, I think I may have that in my dirty laundry pockets.!

Jamie

But not as as many as they sold of other models in the past.

eyond

n them

troit

g of

erious

ks in

nd

ch has

ve

ly a

ss.

LOL

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

If you take apart one of those Chinese bicycle safety flashers, the LEDs are paralleled. I guess there is no current hogging in China.

Actually, it's better than okay- the LEDs have a lot of internal resistance so the current shares reasonably well, especially in terms of visual brightness.

Not every design has to be gold-plated!

The Intersil (originally, now second sourced by many others) ICL7107 meter chip (static LED drive with MOSFETs) just uses a double-size transistor for the leading '1' and requires that the LEDs be paralleled... and the end user there can compare the two LED brightnesses side-by-side. BTW, that part is being used less these days because those dumb Chinese have come up with a less costly circuit that uses a cheap microcontroller and a handful of jellybean parts.

...and the temperature coefficient is in the right direction.

I still get the willies using the internal resistance of a battery as the ballast.

Eyeballing the datasheet, what LEDs are paralleled?

I know the designer of the LED section of that chip. I know the A/D guy too, hot that it's relevant. They did what they had to do using the technology available. When Maxim decided to do a new LED driver, that double transistor scheme was scraped for current mirrored PMOS drivers. Instrument designers don't like their display to look like shit. Instrument customers don't get the warm and fuzzy when digits are different brightness, or worse yet there is ghosting. The subsequent low EMI LED driver actually had a reference to make the current more predictable. [The Maxim parts set segment current with an external resistor.[

Clearly I don't have a future designing Chinese shit. Oh well, I guess that rules out doing a contract with Apple.

Segments b & c of the leading digit (or called segments a and b if it's a dedicated '-1'type).

And an excellent job they did too.. except maybe the AZ cap arrangement- hard to get the noise down to the 15uV spec. But they probably didn't have the real estate to do it digitally back then. The on-chip sorta-reference is pretty lame, but again they probably didn't have the technology.. heck, it's still being used 30+ years later- I started with the gold and white ceramic parts.. and we used serious $$ worth of the plastic parts.

No, we sure don't. But even the original and all the copies (IIRC one of the first clones was Taiwan's ERSO (a government owned semi company at the time).. didn't have any problems. Visible variation due to different LED bins was more of a problem (and TI Korea's quality control was miserable at the time).

LEDs have improved so much that the 8mA current the original chip used would make blindingly bright displays. Some instruments actually PWM it with a 555 or whatever to reduce power consumption.

Or almost any other consumer product company then. 8-(

Best regards, Spehro Pefhany

AIUI, the two LEDs in the most significant digit - the leading '1'.

Right, with the double-sized transistor doubling the mirrored current for the two LEDs in the leading '1'.