free current shunts

John Larkin a écrit :

I once needed a very low inductance shunt for nulling (SMPS assisting) the DC/LF current out of a HF power amplifier. The shunt was a simple

0.5"x1" track with a really thin first dielectric layer. The shunt had to cope with up to 140A RMS HF current and was heat sinked to the liquid cooled HF amp heat sink through a layer of thermal gap pad. The shunt dissipated IIRC 12W on its own.

Once, the prod guys managed to forget to peel off the gap pad protective sheet before assembly. Some Joergish noise and smoke happened instantly... and it took ten $4K boards for them to react! At least the customer knew that I didn't over engineered it :-)

Now, I sometimes place small tracks like that to conveniently add some zeros for free on the output bulk cap of some high perf power supplies I do. It once scared the hell out of the customer's staff at a design review :-)

--
Thanks,
Fred.

If copper thickness variation from lot to lot is a concern you could lay some track in a less occupied area of the board (I know, I know, such areas hardly exist), send some DC from a fairly precise current source through it, measure the voltage drop and then you can calibrate out the thickness deviation. Of course the trace should not be too skinny so that under-etch and stuff aren't too much of a concern.

--
Regards, Joerg

http://www.analogconsultants.com/

That's what the right-most unconnected shunt is for. I can run an amp through the two outer test points and measure the voltage drop across the inner ones.

On the main board, I've got a pair of SMA footprints I can load. Between the connectors a supposedly 50 ohm trace will tour all the layers. I can TDR that and see how well the stackup and impedance calcs worked.

John

An amp through a short strip is a lot but you can pulse it. Doing the measurement for a millisecond every few seconds is probably going to be enough.

That's ok for checking but in production things can change with the next batch of boards.

--
Regards, Joerg

http://www.analogconsultants.com/

The standard model M1A1 current shunt is a paper clip.

--
Many thanks,

Don Lancaster                          voice phone: (928)428-4073
Synergetics   3860 West First Street   Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml   email: don@tinaja.com

Please visit my GURU's LAIR web site at http://www.tinaja.com

We make our own current shunts for the feedback loops of NMR and MRI gradient coil drivers. We use manganin sheet, punched or photoetched into shapes, annealed afterwards. We epoxy them to anodized aluminum heat sinks. Our problem is that we need to make pulses that settle to PPMs in 100 us or so, and bringing the shunt anywhere near a heat sink (which we have to do for thermal reasons) causes horrible eddy-current effects, really messing up the step response of the shunt. To make things worse, we can't afford any net loop area, or we'll pick up 60 Hz hum from stray fields. Having a 1 PPM 60 or 120 Hz line in an NMR spectrum is like a restaurant having roaches in the soup.

Hard anodize is a pretty good insulator and is thermally way, way better than any sil-pad type thing

John

Did you calculate how crappy copper is for this purpuse due to its temp. coefficient? I had this idea a couple of years ago but canned it quickly after I did the math.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

About 0.4% per degree C. I knew that.

I'm only interested in measuring about how much current each supply is delivering. 5, even 10% accuracy would be OK.

The FPGA we're using can theoretically pull up to 4 amps of core current, but that's probably an insanely packed design running at a ludicrous clock speed. The one on the eval board is just barely warm to the touch. It will be interesting to measure supply current and see where we are. Knowing stuff like this is great for doing future projects.

One could do tricks to measure the current accurately, even with these shunts, but we don't care that much on this project.

John

Until recently I had a resistor array with very good RF properties, no eddy currents. It was in a metal bucket full of SAE10 tractor oil. Perfectly fine for about 30 years. Last year the bucket rusted out and it started making pee-pee. Get a new bucket? Hmm ... I decided to scrap it since I hadn't used it in five years or so. _Then_ last week I would have really needed it. Hurumph!

I guess the other alternative is BeO but, of course, nobody wants to work with that stuff for obvious reasons. It's dust is highly toxic.

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

Paging Rich the Dreaded Grammarian...

Oh, yeah, oops ...

Your rite, you're remarque is kerrect :-)

--
Regards, Joerg

http://www.analogconsultants.com/

I've read about copper resistors being used for successfully is ballasts in big arrays of power devices. You don't care so much about the exact value, and the large positive tempco is mildly helpful in improving the current sharing.

Cheers

Phil Hobbs

On a circuit board you could easily compensate for this by measuring the temperature close to the shunt resistor. If it has to be cheap maybe use a BAV99, a resistor and pipe that to an ADC port of a uC.

--
Regards, Joerg

http://www.analogconsultants.com/

Have you considered either kerfing the heatsink (parallel kerfs to the current-carrying direction), or making a heatsink of laminations of anodized aluminum plates? It should keep the eddy currents down, and loss of heat transfer in the direction that crosses the kerfs/laminae is not going to hurt the cooling process.

Oops, that should be 'kerfs NORMAL to the current-carrying direction'. I should wait till the coffee sinks in before I post.

In this case, I think you're looking for the Apostrophe Police. Grammatically, the post might be a bit colloquial, but there aren't any really glaring grammatical errors.

I'm about to give up on my apostrophe crusade - it's like trying to hold back a tsunami with a feather. )-;

Thanks, Rich

Dave Barry is also good at this - I once heard him (on teevee) remark that there are so many ways of spelling there, their, they're, and so on that just about any random string of letters will work, or something like that.

just for fun:

Cheers! Rich

I've always enjoyed Dave Barry's columns. Flew American Airlines a lot and his columns were in every one of their magazines.

--
Regards, Joerg

http://www.analogconsultants.com/

!