Small-scale BGA soldering

It gives you a rule of thumb for rough time/frequency domain conversion. As in, a 120 ps edge probably doesn't need an 8-GHz amplifier.

Cheers

Phil Hobbs

Reply to
pcdhobbs
Loading thread data ...

Still no justification.

How about this:

Take a switching transitor with zero switching time.

Add a single pole low pass filter to produce the observed risetime.

Calculate the resulting bandwidth.

Crude, but it at least has some justification.

Reply to
Steve Wilson

But I've tried, and I've tried, and I've tried, woo hoo hoo...

;)

Cheers

Phil Hobbs

Reply to
pcdhobbs

Not sure. The round ones are plated though the board, the square ones are only on the surface. I picked the round ones because that's what I had on hand.

The customer paid a contract engineering house for a succession of pretty prototypes that wouldn't meet cost and performance specs at the same time.

The engineering house saved the sampler design for last, charging the customer for all of the ancillary engineering for power supplies, firmware, basically the whole project except for the sampler.

The proto pictured validated Phil's simulation work on a sampler design. His work was also done on a very tight timeline. The customer had already been burned and was not feeling particularly patient.

The proto validated that the design met specs. No one was concerned about how it looked.

--
Best Regards, 

ChesterW
 Click to see the full signature
Reply to
ChesterW

I use Liquid Tin for the same purpose.

Reply to
John S

LOL

Reply to
Steve Wilson

I just found out that nickel solders beautifully. You just need the right kind of flux.

First, I tried with a very weak flux from eBay. Didn't work.

Next, I tried a vey old flux I got from somewhere that works real good. It is not acidic, but it leaves a residue that takes isoprop to clean. I suspect it is rosin flux. Probably Kester.

It works beautifully on nickel.

Reply to
Steve Wilson

On a sunny day (Sun, 15 Oct 2017 19:23:26 GMT) it happened Steve Wilson wrote in :

Agreed. Yes, that last one if fun, I remember doing measurements in the old TTL days and showing the horrible waveforms to somebody.... CMOS was much better (CD4000 series), but still..

Reply to
Jan Panteltje

On a sunny day (Sun, 15 Oct 2017 13:54:55 -0700 (PDT)) it happened snipped-for-privacy@gmail.com wrote in :

One problem is that scope V amplifiers do not always (and these days often do not) form a 1 pole LP.

Reply to
Jan Panteltje

On a sunny day (Sun, 15 Oct 2017 12:20:52 -0700) it happened John Larkin wrote in :

Yes, OK, no problem. I was thinking(tm) that all this is from the perspective of scope manufacturers. Of course it sounds great to say: OK we did reverse Fourier and made a pulse and guess what, we only need nn sinewaves, so from this we assure you that there will be no higher frequencies in the pulse you measure and see on the scope.

That is the same as saying, with a DC multimeter, measuring 10 V DC with on it superimposed 5 Vpp AC, 'as soon as the needle no longer vibrates, then it is true DC your measure.' At lower frequencies the needle will wiggle. So it _IS_ sales talk, salesmen lies,.

To figure out bandwidth you need a spectrum analyzer, not a scope.

From my side, pointing the finger at myself, indeed there need not be the 8.3 GHz, you could probably synthesize a waveform where that spectral component is not present.

Reply to
Jan Panteltje

On a sunny day (Sun, 15 Oct 2017 20:36:10 GMT) it happened Steve Wilson wrote in :

OK

They also conclude that you need suficient scope bandwidth...

Good advice.

Nice examples!

Anyways see my reply to JL just now, use a spectrum analyzer.

Reply to
Jan Panteltje

On a sunny day (Sun, 15 Oct 2017 19:19:09 -0500) it happened ChesterW wrote in :

OK. I only use round ones on the surface:

formatting link

Reply to
Jan Panteltje

Yeah, you'd hope not, because if they did there'd be one stage that was limiting the bandwidth very badly all by itself. Gaussians and Bessels and their ilk have the same rise time vs BW, though, within 10% or thereabouts.

My TDS 744A actually did have nearly a 1-pole rolloff, because Tek used caps across the differential outputs from the front end to reduce the BW from the 784A's 1 GHz. Now it's more Gaussian. :)

Cheers

Phil Hobbs

Reply to
pcdhobbs

One of the sadder days of my professional career was back in about 2004, when on a conference call with a bunch of _Tek_factory_engineers_ I had to explain that a scope lives and dies by its step response.

They gave me some cockamamie story about jitter going up due to the slow rolloff of a Gaussian. I pointed out that I could fix that 100% by averaging only two acquisitions, whereas there was nothing I could do about the gross overshoot and pre-shoot of what they were giving me. _Seven percent_ overshoot on a $70k scope, sheesh.

Hopefully my 1180Xes and TDS 694C last awhile. I have a small tube of the trigger ASICs for the 694C, which are what usually fail.

Cheers

Phil Hobbs

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

On a sunny day (Mon, 16 Oct 2017 05:50:51 -0700 (PDT)) it happened snipped-for-privacy@gmail.com wrote in :

Part of the calibration of some digital / analog Teks was bending the wires to the V deflection plates. That capacitance combined with the driver output resistance forming an important pole. Cannot remember the type number.

Reply to
Jan Panteltje

More like 1.5 pF, which is 50 ohms at 2 GHz all by itself. (That's the parallel-plate value, which in this case will be low by maybe a quarter on account of fringing.)

Now iterate over maybe 5 of these connected to the signal nodes. Gets ugly pretty fast. Dead bug is way, way better, even with through-hole parts. A 1/10W leaded resistor isn't very different from an 0603 except that it's a lot easier to solder. ;)

You don't in any case, as long as it doesn't have slots in it.

Cheers

Phil Hobbs

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

This sort of thing is common these days:

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Yikes. That POS wouldn't last long round here, for sure. What the hell would you use it for?

Cheers

Phil Hobbs

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

Looking at slow signals?

Why can't people run things like that through some DSP to clean it up? Maybe because they would lose some of their bandwidth claims?

We evaluated a bunch of fast scopes; all rang to various degrees. One R+S was about that bad too.

If we are doing automated cross-time or risetime or mask testing on production gear, we can (or, we have to) live with a bit of ringing. That's what people are selling.

I'd rather have a less aggressive input filter and a little more aliasing.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

You said it. Fortunately there are lots of boat anchors still around.

Cheers

Phil Hobbs

Reply to
pcdhobbs

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.