ECG003

Strange, and I thought ECG parts were bought up by NTE. ;-)

Tim

-- Deep Friar: a very philosophical monk. Website:

No, Philips bought them from Sylvania. Then NuTone electronics bought the assets from Philips. Now it's just their brand of service aids.

Do you remember these:

GE GE series? Motorola HEP series? RCA SK series? Workman WEP series?

There were others. Do any of you remember any other replacement series?

InGaP

It would have better frequency response if it was in a better package. While SOT-89 is popular, it isn't too good of a microwave package. I wish they'd re-package it.

I use the ECG003 and even more of its sibling, the ECG008. That +42 dBm IP3 @ 100 MHz for the ECG008, and its response down to DC, is hard to find in other parts of similar size. The bias voltage of 7.3 V is on the high side for this sort of part.

If you like these gain blocks, you might also like some of the RFMD "SGA" SiGe HBT gain blocks. I use those too. (Stanford Microdevices to Sirenza to RFMD)

lol.

The original maker of this "ECG" series of MMICs was EiC.

WJ bought EiC. Triquint bought WJ.

I just tried an SBB-4089Z, from RFMD. It's unusual in that it's flat to 6 GHz, whereas most MMICS droop badly over their spec'd frequency range. Time-domain step response is very nice, no slow drool like most MMICS. Rise/fall times are like 50 ps, consistant with 6 GHz bandwidth. It's a bipolar darlington, but doesn't seem to have any saturation recovery delay.

This one is $2.24 at 1K, a remarkable amount of GBW per buck. Digikey has them.

Yeah, it is active bias so does not have response to DC. If you don't care about S11 & S22 and gain below about 50 MHz, then it is a good part.

If you need wide band response, you'll spend more on the conical inductor from Coilcraft. hah hah.

Check out the NBB-400 too. They claim a 3 dB BW of 7 GHz. Not a SOT

-89, but easy to handle for any production environment. It does not have quite the power capability of some of these others.

To get really flat response, it means a TWA, and those are expensive.

Since it's only AC coupled, eventually it has to drool...?

Have any screenshots, by chance, of what your typical "slow drooling" amps look like?

I once attempted to take a high-speed AC-coupled amp like this and then use a slower parallel path to "fill in" from DC to a MHz or so. Despite a lot of effort, it never did work very well -- I couldn't get, e.g., a step response with

I like the big ground pad, electrically and thermally. And SOT-89 is cheap.

Take a look at SBB-4089Z.

Sometimes I cheat and pull the input pin up or down to shift the output voltage, especially if I'm amplifying pulses of one polarity or the other.

I've used the SGA3586. Very low noise. Most MMICS don't really have 50 ohm input impedances (they are typically lower, some as low as 30 ohms) but this one does. You can tune the device current and nail 50 ohms exactly.

The 3586 is fragile, easily zapped with input spikes.

That's the low end. I meant the fast step response.

Sorry, I didn't photograph it. The step response of the ECG003 is about a 100 ps linear slope up to half of ultimate amplitude, then about a 500 ps tau exponential the rest of the way. Adding a 50r resistor paralleled by 10 pF in series with the input fixes it pretty well.

That's not bad at all.

Yeah... I would have been happy with even, say, 3%, but it's amazing how just a little 1dB dip in a relatively narrow frequency band seems to do almost the 10% damage that a 1dB dip across the entire frequency spectrum would in changing the step response's overall amplitude by 10%.

I was surprised to learn recently that Tek high-voltage differential probes (e.g., the P5200 series) have step responses with a whopping ~20% overshoot. ...but I also learned that LeCroy and Agilent are likely no better (I just haven't tested any of theirs). Hmm...

---Joel