BJT Pair for Push-Pull, 1W, 33V, 200mA, 1MHz (& maybe no heatsink)?

out

description.

then

would

I take it you can't access binaries groups ? I'd post something there otherwise.

One thing puzzles me. If this is either current or voltage output, how is the output level and impedance configured ?

Graham

What does *that* do ????

Graham

"Peps" it up ;-) Sustained operation burns valves.

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |

Ouch ! Not as bad as sugar in the short term at least.

Graham

Tom, For some reas> I think I'd [...] make some kind of direct current source with OpAmps

That's exactly the approach I was taking with my request last week for a voltage-controlled current source, which Ban answered. It's for a fully automated curve tracer like Tom's, but with minimal front panel - just a 2x16 LCD and a "Test" button. When you hook up a device and press "Test", an internal MCU goes through a series of probes to work out what's connected and with what polarity, then goes on to characterise it and spit out curve data on a serial or USB port to display on a PC.

Hence I was looking for a current source with up to perhaps +-15V compliance that'll track from a few uA to tens of mA. I think that's a better approach for a tester like this, as most devices will fail due to excess current, not voltage.

Clifford Heath ( snipped-for-privacy@gmail.com)

John,

That's very interesting.

Tom

be

prevailing

out

description.

somewhere, then

would

otherwise.

output level and impedance

Graham,

Thanks again.

That's correct. I haven't found any free way to access the abse group. I'll check as to whether or not my ISP will agree to subscribe to it.

But, you could email something to me. My regular email is still "hosed". So I'm currently reachable at: tomgootee at yahoo dot com. I could also then post it on a webpage, if you'd like.

Regarding the output level and impedance configuration: With the Howland topology, at least, the 100 Ohm current-sense resistor (in series, between the output and the load), and the 100 Ohm resistor that could be manually switched-in, to ground, from just before the current-sense resistor, gave excellent output-levels' accuracies, FOR the intended load conditions, i.e., when using current-source mode (i.e. without the 100 Ohms to ground) for low-impedance (e.g. BJT bases) and short-circuit loads, and when using voltage-source mode (i.e. WITH the 100 Ohms to ground) for high-impedance (e.g. FET gates) and open-circuit loads.

That may be a wholly-wrong approach. I hope that it was implicit, in my even asking about all of this, that I don't (yet) consider myself to be very experienced, or even good at, analog circuit design and analysis, and pretty-much continuously wonder (and would like to be told) if I'm making any serious blunders, maybe-especially regarding the very basic things, such as my overall approaches to design solutions. (I do hate having to possibly expose my ignorance in public. But, if I must, I must.)

And, in case I haven't mentioned it lately, and just so everyone here knows that they might need to "dumb-down" (somewhat, at least), their responses to me: I DO have a BSEE degree (very near the top of my class, actually, at Purdue, 1978), and even took quite a few graduate-level courses in EE and math. *HOWEVER*, after only three or four years of doing (cooool) guidance and control work on fighter-aircraft programs (And I still like to brag that I made a major contribution to an air-to-air automatic-weapon-delivery system; Great fun!), I spent most of the rest of my life writing software, letting much of my EE knowledge rust or evaporate, and have only been really-seriously trying to learn and gain experience at circuit design for the last several years. I HAVE been working quite diligently at it, but still accutely sense what feels like a gaping hole where a career's-worth of experience would be so extremely useful to have, now. This group (and its archive at

Thanks again.

Humble regards,

Tom

You devil, you!

This same thing happened to this vehicle, once before, just a month or so ago, when I had parked in the same place, one evening. It "seemed" like water contamination. I suspected tampering, since it was running perfectly before being parked. It idled very roughly, ran poorly, with "random misfirings". I later filled the then almost-empty tank with 93-octane gas (I usually use 87) and the problem was almost-instantly gone.

This time, after parking in the same place again (doh!), the symtpoms were similar but more severe. I left the vehicle near there, at a friend's place, after having driven a mile or so and deciding it was running way too poorly to drive, and also thinking it might be best to not run it with whatever contaminant might be in the gas.

I went back, later (yesterday), and drained the tank (almost), and then filled it with 93-octane again. But, this time, "no joy", although it seemed a LITTLE better, and more-or-less drivable. So, since the "Service Engine Soon" light was still on, I went to a nearby Autozone store, where they will read the computer codes for you for free (if that light is on). Not too much info resulted, though. Got three codes, two of which were "Random ... misfirings", and one of which was something like "O2 (exhaust) heat sensor unit 1 bank 1". So I dumped in a bottle of "highly recommended" fuel-injector cleaner and drove it home.

It idles roughly, but seems pretty-much OK at higher RPMs. When backing out of the garage, it sounds like it's hitting on about one cylinder (dunk-dunk-dunk-dunk-....). BTW, this is a 350 cu in Vortec V8 engine, in a 2000 Chevy Tahoe, with 97000 miles on it.

If it doesn't get any worse, I guess I'll probably just drive it until most of this tank of gas is gone and then see what I need to do. I hate most repair shops, especially the car dealers'. But, it's probably due for a "real" tune-up, anyway. ("Where's the K-Y?")

Tom

(1) Get a locking gas cap... I got one for my truck when fuel prices began to rise. (The Q45 has a locking fuel door.)

(2) If it doesn't clear up with the fuel injection treatment it may well be a sticky "air door" (in the air flow path, controls the injector duty cycle). I had that happen last year. If you can't fix it yourself, expect about a $1K replacement cost :-(

...Jim Thompson

Clifford,

Sorry you can't get to fullnet.com (my ISP; my stuff is at

Possibly unrelated: Once a year, or so, I do have a problem with sending emails, when my ISP (fullnet.com) gets blacklisted as being an "open relay", or something like that, on some spam-site listing. I sort-of know the owner of Fullnet, and talked to him about it, once. He truly hates spam and said he had everything buttoned-up pretty thoroughly, and didn't know why that was happening.

At any rate, I kind of like the idea of your automatic probe/sense & test system. But I also like to have detailed control of both currents and voltages, with as many knobs and switches as necessary. I think it might be nice to have both ways available.

And I, too, used to worry about the cost of the components, usually first and foremost; but not as much, any more, or, at least, not if it interferes overly-much with what I want to accomplish. Just the NKK rotary switches that I now use (MRK series), alone, are something like $10 each. And I have five of those on the front panel. And parts like REF102B, LM394, AD633JN, LT1364, and even OP275G aren't exactly cheap. I also use QUITE a few 0.1% resistors, and those expensive Caddock Power Film Resistors (about six of their MP930 series in each unit).

One thing that does cause me pain is when I see what they want for some of the instrument cases/cabinets that I'd like to consider using. And don't even get me started on the prices of small cooling fans.

I'll have to go find that recent thread you mentioned, about VCCSs. But were you saying that you only control current, and not voltage?

Tom

You might want to check any fuel filter you have for left over gunk.

Robert

I'm familiar with that problem, but that's not it - I couldn't get a DNS record for your site (today it works). You should tell the owner that he needs to get a more reliable DNS service. Or maybe it's his fault; he seems to be running his own DNS server, perhaps it was down?

Why not through a PC-based front panel though?

Well, I figure that even a good oscilloscope only has 8 bits of resolution, so there's little point in making measurements to greater accuracy if you're only going to display them on an o-scope. I know you *can* do more than that, but how often is it needed? Normally when you want to characterise a component, you plan to use several of them, so have to design for variations anyhow.

Yes, hardware is expensive - or maybe it's just that silicon is so relatively cheap, especially if it's in a commodity PC :-).

All you'd have to do it switch in a precision resistor and you'd have a voltage source with a defined impedance...

My thought in using a current source is to inject currents into two points and measure the voltage from those to a third ground point. Gives you the same information as voltage sources, but doesn't require additional current limiting, and uses ADCs in voltage mode with nothing more than a high-impedance buffer needed.

Clifford Heath.

Thanks, Robert. I WILL check the fuel filter.

I hate car trouble.

Hmmm. While sitting here thinking about that, I remembered what's pretty-much a "good luck" car-repair story, at least from one point of view: About thirty years ago, someone I knew had acquired an antique Mercedes, and was having it towed to a specialty place about an hour from his house, to have the body restored (which sorely needed it). On the way there, a truck ran into the towtruck and the car, causinging fairly-significant damage to the body of the car, but not much else. The truck's or towtruck's insurance then paid for all of the body work.

Thanks again, Robert.

Tom Gootee

= Why not through a PC-based front panel though?

(Just "off the top of my head":)

I think that that could be great, especially as an adjunct. But, if it were the ONLY means of controlling the Curve Tracer, it would make the instrument (totally) dependent on the PC (and its software and/or operating system), which may not be desirable, in many applications. It would also tend to limit the instrument's overall availability, reliability, portability, and long-term viability, in general. It seems like it could then also limit the size of the potential market, as well.

Of course, if the instrument already *REQUIRED* a PC, then you definitely might as well also have that.

On the plus side, *requiring* a PC, for an instrument to be able to function, might: remove some significant portion of the hardware requirement from the instrument and shift it into the PC, which could also already be in a potential customer's inventory, possibly lowering their overall acquisition cost (and/or raising the manufacturer's potential profit, or, simply increasing the maximum potential "bang for your buck" of the (non-PC portion of the) system), and might also even improve the reliability and maintainability of the (then-simplified) specialized instrument portion of the overall system, as well as provide easier and cheaper means for implentation of other potential features and benefits that might use parts of the PC portion of the system.

(Whew. I'm sure there was a much more concise or elegant way to say all of that.)

I am also considering simply putting my whole Curve Tracer ** INSIDE THE CASE OF A PC **, and mounting the front panel controls, etc, in some convenient place, perhaps on empty drive bays' front plates, and then connecting interfacing to the PC internally (probably either a plug-in card (e.g. PCI) or a USB port, or... whatever). (Ever notice how very closely the early Tektronix digital scopes resembled a PC with an oscope just added in?)

That approach would also simultaneously solve the "too-expensive instrument cases" problem I mentioned earlier (basically, making the case FREE), and, actually, seems like a pretty good approach to building almost ANY type of benchtop-sized piece of equipment that needs any processor, display, networking, etc, etc, functionality that a PC could provide (and possibly power rails, too).

-----------------------------------------------------------------------------

Actually, one of my BEST ideas, YET, *might* be THIS one:

BUILD INSTRUMENTS AS "PLUG-INS", WHICH WOULD FIT INTO STANDARD PC DRIVE BAYS!

(Or, if an instrument used only the PC's *SCREEN* as its "instrument panel", it could even be built into a standard laptop plug-in card, or into a laptop floppy or CD-drive form factor.)

----------------------------------------------------------------------------

Regards,

Tom Gootee

tomg AT fullnet dot com

----------------------------- As my friend Jay says: "The tempus it do fugit."

-----------------------------

Couldn't I just convert the Howland-based circuit into a more-typical "opamp with booster amp" type of circuit, basically by removing the feedback loop from the output to the opamp's + input, and, of course, changing the value of the resistance in the feedback loop that goes to the opamp's - input (and making a few other obvious, more-minor changes)?

...Tom Gootee

Hi Joerg,

Thanks.

I do have D1 and D2 thermally coupled to their respective transistors.

What size emitter resistors might be reasonable?

Regards,

Tom

then

Quite

Graham,

What sort of emitter resistances would be reasonable? And what's the best way (if there is a way) to increase the output level, which will have been lowered by raising the emitter resistors' values?

I did try a non-Howland topology in Spice, by removing the positive feedback loop and changing the neg loop's resistance to 10K (along with the opamp's input resistor), and connecting the opamp's + input directly to ground, and disconnecting the offset-correction injection.

With only three-Ohm emitter resistors, I get the output tracking the input voltage until the input voltage exceeds about 14.61 volts. The maximum output voltage/current level is about 14.61V/145.9 mA, for a

Also pressure regulators.

...Jim Thompson

Jim,

Thanks.

I guess I'm not familiar-enough with this vehicle, or automotive systems in general, to even know if there are any of those. I guess I'd better do a little research, and maybe try to trace the fuel system from the tank forward and identify any components that I might need to check.

Tom