Newer Model Instrumentation Amp

On 20 Jun., 19:14, John Larkin wrote: snip

That is DSP 101 and kinda obvious when you think about it, for a filter to have any effect at 1Hz it needs to have "memory" it that range. with a 500kHz samplerate that either means a very high order or a very high precision

standard approach would be decimate the signal to a more resonable sample rate

-Lasse

You'd think that 16 bit data and 1e6:1 sample/cutoff ratio could be accomplished with roughly 36 bit math.

That's one approach: build a chain of filters, clocked at declining rates, and use the early ones to take the stress off the later ones. Switch filters in or out as needed. We could do that, but it's harder to explain to the customers. It's easier to model the product as a lowpass filter followed by a 500 KHz ADC.

The classic butterfly explodes violently as the clock:cutoff ratio goes up. Numbers have to subtract to make tiny residuals, and stage gains get absurd. Instead we used a double-integrator filter, essentially a digital simulation of an analog state-variable filter. That gets into trouble much more linearly than the butterfly, and each

2nd order section has a gain of exactly 1. It works using MAC blocks using 18x18 multiplies and a sort of coarse barrel shifter into a 48 bit accumulator.

John

K is the ANSI/Mil designation for a relay. OK, I'm old fashioned and don't just make up stuff like LED2, TR5, RLY12, CON2, RV7, RN8, ZEN15, or other abominations.

John

How does that help? Idss is only a few mA max, and the resistor just adds to the Johnson noise. And takes up space.

That does make sense if you use the bigger parts and can't stand full Idss.

John

It helps if you want to protect against someone connecting to say 230Vac, and leaving the connection for a while before discovering their error. We need to protect the sot-23s.

--
 Thanks,
    - Win

Ok. 230 AC peaks at 325. Half cycle, the average voltage is 104 across each fet. Assuming 2 mA, that's 208 mW. A SOT-23 should be OK with that, but one more resistor could pull that down some.

In my case, dissipation is a bit lower since the other sides of the fets are whacking the ESD diodes in the opamps, roughly +-18 volts.

John

Dunno about that! I was just combing though some BOMs looking for a couple nickels to squeeze and just happened to notice that we're using a $7 cap (Case-D 220uF 20% 16V) on one product. I have no clue why we're using that cap because we have a similar cap (Case-D 220uF *10%* 16V) that we pay $.70 for. The relays on the $7 cap board are only a buck or two (I'm not looking at that board, so don't remember exactly).

I've never seen anything other than 'K' used. I'm with John, RL, REL, and all other bastardizations are just that.

In my ideal world this is the sort of thing the purchasing guys would notice. "Hey Mr. Engineer -- is there something special about this cap that makes it worth 10x this seemingly superior one over here?"

Although in my ideal world the engineers at some point went over the BOM to do a little sanity checking on prices as well, and would also notice it. Before hiring you, I mean, since I imagine it's a product that existed before you were hired. :-)

If one's not going to use 'K', 'RLY' is a lot better than RL or REL.

Abbreviations are kinda strange, though. We occasionally refer to spectrum analyzers as "SpecAns" for short, but I have a radio that refers to them as "SpeAnas." Weird...

---Joel

With lead times going out into forever, we're just lucky they've not had any purchasing disasters this year. OTOH, we have inventory of some components that go out three years, or more.

Actually it did. ;-) It's used in the football version of the unit you played with and came out a year before. The good news is that there is only one per board, the build numbers are in the low tens per year, and the product has *reallY* fat margins (5-10x that of what you saw).

The real problem is that engineering never saw prices after the original look-see in DigiKey during design. I asked our admin to dump the purchasing database by part number and I'm going through the BOM manually (actually, I placing the cost info into the schematic properties pages). I'm trying to get the IT folks to feed back the purchasing database into our engineering database at least monthly, but so far no one is very interested.

I suspect the kids have never used them so make things up as they go. No reason to use any such abortions. Relays get 'K', period. ...might just as well use that commie 49R9 and 3K9 crap. ;-)

That's the problem with such abbreviations. Everyone does it differently.

As far as the "SpecAns" goes, how about "FreqDomScope", or better, "that gizmo over there"?

That sounds like a pretty good strategy.

I didn't realize there were many markets that supported fatter margains than the pro audio guys. Wow!

Well, maybe jewelry or sporting goods...

That's a useful idea, and if it's automated you could easily run it, e.g., every night or so. We're still just working on getting ORCAD CIS tied into the Big Expensive MRP program; purportedly all the hooks and configuration bits already exist, but someone needs to sit down and define the actual fields we want to use, (more importantly) standardized methods of filling in those fields, etc. -- and then actual start filling in all that data for the parts that already exist in it.

Yep, I think you're right.

We'll occasionally refer to, e.g., "the 8594" or "the 9020a," but if someone isn't remembering the particular model number, it turns into "the really expensive specan" or "the el-cheapo specan." :-)

---Joel

[...]

Commie? That's different, got a schematic here but my PC can make those na'sdarovje letters :-)

I am sometimes tempted to use 4R99. Especially after another call from someone where it went *PHOOMP* .. *SPLAT*. "I thought it's 4.99ohms" ... "No, that must have been a speck of toner from your printer, it is 499ohms".

"them thar thang over yonder" (for the guys from the west).

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Is that aluminum? $7 is amazing. We stock a 220u, 25v that's 19 cents.

I was just reviewing my ADC board and noticed a half dozen 47 uF 25V tantalum caps, $2 each. We have a 33 uF in stock for 18 cents. Good enough. Some of the regulators, like 337s and 1117s want tantalums on their outputs.

John

I suppose with amateur radio guys "panadapter" counts too...

But the really tranditional guys would never connect that to the IF, they'd have a Q-multiplier on there. With a tube, of course :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Not really. Banks hate it.

Multiply an 87000 seat stadium by a $50 ticket, with free labor and come back to me. ;-)

Only the layout guy has CIS. What good it is for him on 16.2 CIS when everyone else in on 15.7 is beyond me. The other hardware designer has his own copy of Orcrap tied to his spice license and I use the floater.

We only have one, and it's in the service department. Well, we do have FFTs on the scopes, witch is usually good enough for what I need.

Yep. Euro-commie crap. ;-)

We do everything on a plotter (HP DesignJet, actually) so the printing is pretty good.

Yep, they only talk like that here when there is a Yankee in the room. ;-)

Tantalum.

We don't use alluminums. They don't like the RoHS process much. Damn Euro-commies.

Hi Keith,

Hmm, good point.

I've always figured that guys manning the hot dog stands are making pretty much minimum wage too.

[Spectrum analyzers]

We have some big boxes that are effectively just fancy RF switch matrices with filters and amplification included and we're always sniffing around them, it seems, with spectrum analyzers trying to improve isolation, make sure things aren't compressing, measuring intermods, etc. For the wireless stuff we design, as long as I've been here it's always been a "roll your own" approach using mixers, amplifiers, filters, switches, etc., and often the scopes remain powered down for days at a time in deference to the spectrum analyzers -- we haven't had a project yet where we felt the wireless radio modem "modules" were the best option, even though they certainly have many compelling features.

Unfortunately, it's becoming harder to find general-purpose radio parts like this -- particularly if you're after low-power operation. If you look at, e.g., I/Q modulators, most of them are around a watt! -- Fine for basestations, not so great for anything strapped to your hip. The iPhones and Razrs and whatnot of the world are driving this sort of thing to a system-of-chip approach...

---Joel

Thank you, John, for posting that real-world schematic. Newbie question: what do R417, R120 and R123 do for the circuit? Calibration of inputs to INA154?

Doug Beeson

--
______________
Douglas Beeson
Java Software Design and Development
Montreal, QC
dbeeson99@videotron.ca - remove the nines