• EEVblog #567 – Precision 1A Current Source

Dave breadboards a precision 1A current source circuit and finds that even the application reference circuit doesn’t work as expected. Showing that you can’t just blindly trust that datasheet application circuit will work first go.
LTC6655
Forum HERE

Be Sociable, Share!

EEVblog #1047 – Solar Roadways FINALLY BUSTED! Colas Wattway

Dave finally puts an end to the idea of Solar Roadways. We have the test ...

• ACM

Hi!
Nice video as usual, I’m designing a pcb with the same reference chip @5V with the npn bjt. Actually I put a 100R restistor between the F pin and the base of the transistor to reduce the bandwidth, I hope it should help!
Anyway now I’m waiting for the next video!
Thank you

• Arnon

why pure resistor redeuce the bandwidth? what abour RC filter?

• Zach

There is always stray capacitance. You’ll notice components are generally low pass filters in addition to whatever their intended purpose is.

• ACM

Hi,
thanks Zach to your firts reply, the components have stray capacitance but in this case we’re working with a transistor and the input as the behaviour of a diode.
Without a resistor, a small rise of the input voltage (base emitter voltage) implies a big variation of the base current and consequently a “proportional” variation of the emitter current.
Adding a resistor this effect is reduced; think how you could calculate the base current
1. Without resistor you’ve to use the input characteristic graph of the transistor;
2. With the resistor you could evaluate the current as: ib = (Vinput – 0.7)/Rbase
Vinput is the voltage applied to the resistor connected to the base.
The two cases are different, think what happen if you double the input voltage!
In the first case the current rises more than twice (that effect could lead to an unstable situation), in the second scenario the current rises in a more “proportional” way.
That’s why I never connect a transistor without a base resistor, anyway if necessary I will mount a 0R resistor (up now never happened).

Sorry for my english.

• Simon

Hi Everyone,

Just a couple of comments (the last one took longer than I expected… sorry!).

Dave talks about the sense output, it is (if I understand the chip correctly) actually an input, it’s just that it happens to be connected to the output node which is why the pin is annotated as a Vout (confusing I think!). I know that Dave knows that, just confusing in the video.

Now the second and (you’ll be glad to hear) my second and last comment…

ACM makes a valid comment about adding a 100R in the base but perhaps this needs a little more explanation…. so here goes(!)..

I have used 2N2222 parts many times, mainly as switches, BUT if you use these devices in a linear fashion (i.e. not switches) you sometimes find that although the configuration might be (as your primary intended) a common emitter or common collector (emitter follower) IF you look at the a.c. circuit you find yourself with a grounded base sort of affair (what a guy from National Semiconductor used to call grounded grid!).

AND the reason he used to use that expression was because “grounded grid” automatically used to suggest (to us old gits) an RF amplifier.

An RF amplifier of this configuration has an un-quenchable (pun intended) thirst to oscillate. If you want to “stop” such oscillation, then insert a base “stopper”.

That’s where I learned about the “base stopper”, a crazy word that really doesn’t explain anything about what’s going on just what it attempts to do.

Just to close, I have to agree with Dave that I wouldn’t expect LF oscillation (especially as it appeared as an application in a datasheet), so, yes, that is strange.

BUT what I will say is that any 2N2222 used in a linear stage (especially using devices pre-2000, yes I have evidence of that)can “take-off” and the base stopper of choice is normally a ferrite bead.

Check out any T&M equipment (80’s to 90’s) using 2N2222 (or 2N2222A or the larger TO3 part 2N2219A etc) in a linear stage and there is a distinct possibility that you will find a ferrite bead (FX1203?) threaded over the base lead.

Well having successfully broken two rules of grammar (using AND & BUT to start a sentence) I shall now shut up!

Simon

• Simon

It’s me again,

A bit of a blunder, as I’m sure you all know the 2N2219(A) (the equivalent of the 2N2222(A) but with higher Pdiss) is actually a TO5 part not a TO3 as I said.

(I’ve just been repairing a stereo amplifier and I’ve got TO3 parts on the brain!)

Simon

• Chris

I’ll have to send this to one of my friends who told me, “If you have noise, just stick a bigger capacitor in there.” Doesn’t always work.

• huh

Those beasts are fast, they easily oscillate at RF frequencies, and a capacitance too far away (or having long legs like that one on the breadboard) becomes a resonant lc circuit that calls for disasters. You probably could “hear” that ic with an AM radio kept at close distance.
The trick is always to stick filter capacitors as close as possible to the in, out and gnd pins.

It would be useful a look at what is being produced on that IC pins with a spectrum analyzer.

• JFA in Montreal

Poor Linear Corp.
Pure P.R. nightmare ! 🙂 🙂 🙂

There was two main frequencies on the scope.

Could that be some noise being picked-up from the surrounding switching P.S. by the very sensitive sense input of the chip? I did not see well enough to notice if the frequency of the oscillations changed significantly upon changing the decoupling cap values, or if it did not move along with the value of the caps.

• Hi Dave,

How about something simpler. Like a LM317AT or a LT3080 set up as a current source. Also I second the ferrite bead. Try type 43 material.

Sam
W3OHM

• Hello Dave – can you tell me where you source from those breakout boards as shown in the LTC6655 video ? [ EEVblog #567 – Precision 1A Current Source ]
Thanks – Michel

• Agent24

You can get SMD breakout boards all over eBay…

• Bill L

I have gotten some from ti.com in the past, \$10US including shipping.
They come with 2 of Samtect TS-132-G-AA headers