4-20mA Precision Source Ideas

That's cute. But at our place we are currently batting 0 for 3 on bootstrapping opamp power rails from their own outputs. They love to oscillate. We eventually got them to work, with difficulty.

John

Yes, I saw your "Larkinesque" version :-)

So your needs are for a non-standard 4-20mA device?

Yes that's right.

Hey, it's unusual for Thompson to issue such compliments.

John

Do you plan on some reference to ensure 4mA at 0V input? ...Jim Thompson

snip

using something like this:

-Lasse

1. Yes

1. It doesn't do what he wants-- it's designed for loop-powered instrumentation.

Best regards, Spehro Pefhany

There are some relatively new 36V op-amps, but they are generally not cheap.

Best regards, Spehro Pefhany

;)

lol, I realised that 10 seconds after I posted, I'm guessing what he needs is what is called type 2 in this:

-Lasse

Right, though it doesn't have the performance hit of the Howland approach for resistor matching. 0.1% resistors are cheap these days, and you can calibrate out any scale error digitally in most modern systems with a single point calibration.

If you _really_ want to get down to just one precision resistor, you can always use a flying capacitor for the level shifting.

;-)

Best regards, Spehro Pefhany

Now now, don't start...

I may breadboard it over the weekend if I have a chance, see if it's stable. I wouldn't trust the models to tell me that!

No, actually I misread you. 0.4-2.0 V is 4-20mA. The DAC itself can output 0-2.5V which would be 0-25mA. It is driving some process indicators which take "4-20mA" inputs, although I gather they can be scaled to various ranges and offsets.

A circuit like this might work for that:

My LM317 model doesn't seem to be very good, though, the current out of IADJ doesn't match the datasheet and the reference voltage is all over the place.

Thanks for the application note - I am a bit ignorant of the standard practice in this area. I think it would be "type 3".

No one has mentioned current sense circuits like LTC6101; with one of these on the high side, and a two-transistor translator, you can monitor the current and control it using any old

You can get over 100V of compliance using high voltage transistors this way.

Ha ha I was just thinking about that. I find all those LTC6943 circuits keep haunting this sort of thing. Jim Williams' personal chip AFAICS.

Yeah, there it is, Figure 17 in AN3. :)

Quite a few still spec'd to +/-22V. I've been looking for them lately, minimum 40V operation to go across 12cell lead-acid batteries, because of the head room the TVS diodes or zeners require. 36V is getting a bit close to the max 30V or 31V seen near end of charge, plus hash and noise and pins from the charger.

Grant.

You didn't give the quiescent current anywhere to go for the '317, of course it's upset by that ;) Hardly needed for a 4-20mA circuit.

Grant.

Only if you make it a two terminal circuit though, classic 4-20mA transmitters are loop powered. Pop in a zener where you need some voltage, and keep quiescent current below 4mA otherwise you get to try 10-50mA, or even 1-5V if the receiver is close by, say in the same cabinet, usually a few options for that industrial stuff.