Quick ESR answer needed

You mean, no trek down to Gordon Biersch's in the evening?

No, too hoppy for my taste. But my attorney's office (he's a board member) is just a couple of blocks from GB. He's in the penthouse of the building at Mission and the Embarcadero, above Boulevard Restaurant. We're sort of his hobby.

Yeah, here it is:

John

Yes, but, what does the 4192 say for ESR at 20 Hz, 100 Hz, 1000 Hz and

Ever tried their Maerzen? Just don't drive after a couple of those.

Nice digs! I guess his hourly rate is quite a bit higher than that of engineers.

Hello John,

That server tries some nasty pop-ups. Ricocheted off here, but others might not be so lucky.

What's the ballpark s.e.d. discount price for that one? I am using the old Philips PM5785B for stuff like this and it's fine for me. But I have a client who may need to generate test pulses with less jitter than the Philips can do. How much oomph can you get out of that optional 50V output? We'd need more than 5V, sometimes a lot more.

I've only been to Boulevard once, and it was fabulous, but I'm scheduled there for my next birthday. They do have $3000 bottles of wine on the list. During the dot.com days, they discovered that the more they charged for a wine, the more people ordered it. In the year

$400 or something like that.

John

No such problems here (Firefox v2.0.0.6).

...Jim Thompson

I guess they must have really suffered when the dot-com bubble popped.

Ouch. My attorney is around $250 but this is in the country, east of Sacramento.

I didn't have any here but you can see the attempts. Only web sites on the more rough-and-tumble side do that.

Oh. I've never seen their popups, just a few static ads. I'm using Firefox.

List price is $3840, with options for OCXO, Ethernet, and stuff like that. Jitter for short delays is around 6 ps RMS, increasing to 10's of ps for long delays. The HV option is $600, and puts 50 volts into

John

I like that classy fluorescent display. Yes, I'd be interested in a manual:

jsc at ieee dot org

(that one is less typing than my biz email address)

Good deal! All I got for my birthday (the 5th) was my wallet disappeared with all my ID, and $150 in cash, which was most of my food money for this month.

When I was in my mid 20's I saw Carol Doda in her swing, and fell in love... at least the lower part of my anatomy did ;-)

...Jim Thompson

John, I'm interested in a manual too. to tschaggelar (at) dplanet (dot) ch

Rene

--=20 Ing.Buero R.Tschaggelar -

I made a business trip to HP in Cupertino, with my bosses, in roughly the early 1970's, and we went to see Carol at the Condor. I didn't care too much for her body, preferring ballerina types myself, but she was actually very talented and her show was a lot of fun. The nasty parts were projected on a screen behind her, because full live nudity wasn't allowed in places that sold drinks. That's still the law here, maybe, I think.

The HP computer group was in a beautiful low glass building in the middle of a peach orchard. I drove by last year and barely recognized it... now surrounded by Burger Kings and car dealers and motels and crap.

John

I hope you are using the term "esr" (I prefer to capitalize it) to mean the same thing the capacitor manufacturers do. They do not use it to mean the "true series resistance" as John Larkin showed in a more complicated circuit model that he and I have been discussing. It means the total losses of the capacitor at a given frequency are lumped together in a single "equivalent" resistor modeled as if it were in series with the capacitor. It is the real part of the *impedance* as measured on your 4192 using the series circuit mode.

You say that "One assumes the esr changes little with frequency", and then you say this is completely consistant with the data. To what data are you referring, and are you using "esr" to mean what capacitor manufacturers mean by it? The plots of ESR vs. frequency provided by the manufacturers all show ESR increasing as frequency decreases into the low audio range.

I would not assume the ESR is (essentially) constant with frequency, but rather that the dissipation factor is more or less constant with frequency, say from 100 Hz to 100 kHz. Since the DF is 1/Q, or resistance/reactance which is R / (1/wC), if it is to remain constant with frequency, then as

At very low frequencies, ESR is not swamped by the capacitive reactance. ESR increases at just about the same rate as reactance. Set display 2 on the 4192 to show Q and see how it varies from 100 Hz to 100 kHz. I did this with a .1 uF MLC and the Q only varied over a range of 60 to 100 as the frequency varied 3 orders of magnitude. If Q is to remain more or less constant, then the ESR must increase just as fast as the capacitive reactance as frequency decreases. And, in fact, this is just what I see when I make the measurement.

The data that I've measured shows that ESR varies inversely with frequency below 100 kHz, typically; it's not constant at all. And this is just what should be expected. In every case where a manufacturer gives plots of ESR with frequency, you'll see that it increases as frequency decreases to low audio.

These conclusions I've given all depend on the definition of ESR as used by the capacitor manufacturers.

"The Phantom"

Cap makers DATA sheets and published marketing claims all refer to minimum impedance values at self resonance as being the " ESR ".

...... Phil

heres one that doesnt:

more importantly, it contains graphs of ESR and impedance vs frequency for a variety of types, and more relevant to this thread, impedance & ESR measurements versus frequency (1kHz - 40MHz) of 25V 10uF (it appears X7R) multilayer ceramic caps and the comparable OSCON part, with both 0V and 20Vdc bias.

Cheers Terry

I don't think capacitor manufacturers are consistant in what they mean by ESR, or how they measure it. I suspect that many of them don't really understand what's going on, and buy some instrument that reports some number, and believe it.

In low impedance circuits, which a lot of us do a lot of the time, the LSR (Larkin Series Resistance) is what matters, and no 0.1 uF ceramic cap is going to have 800 ohms of LSR.

John