1. Home
  2. Electronics Design
  3. 1 mA Current Source Tests

1 mA Current Source Tests

Posted on ABSE with attachments.

I'm using LM334s for 1 mA sources I need. A while ago I posted, asking if there is a dual 1 mA source chip. No joy - but the ensuing discussion whetted my appetite to see if a proposed sink circuit using an LMV431B could be converted to a source that would perform as well as the LM334 circuit. (The tempco circuit on the

334 datasheet).

For clarity: source, not sink. The load goes between source and ground.

I've attached the original proposed sink circuit, and the test jig circuit I used for the following comparison. The same 2181 ohm resistor was switched to the test points for the Vr measurement. Temperature was raised over the course of 1/2 hour, and measured with a digital temperature meter. The voltage was measured with a Fluke 87 V

*Vcc = 10.00* ************** LMV431B *** LM334 **** Op Amp ** *************** TP1 **** TP2 **** TP3 *** Temperature *** Vr **** Vr **** Vr *** *************** ***** **** ***** **** ***** *** *** 72 F **** 2.181 **** 2.181 **** 2.182 *** *** 74 F **** 2.178 **** 2.179 **** 2.181 *** *** 78 F **** 2.174 **** 2.179 **** 2.181 *** *** 82 F **** 2.173 **** 2.179 **** 2.181 *** *** 86 F **** 2.172 **** 2.180 **** 2.180 *** *** 90 F **** 2.171 **** 2.181 **** 2.181 *** *** 95 F **** 2.169 **** 2.184 **** 2.181 *** *** 100F **** 2.166 **** 2.185 **** 2.187 *** *** 105F **** 2.164 **** 2.188 **** 2.187 *** *** 110F **** 2.162 **** 2.188 **** 2.188 *** *************** ----- **** ----- **** ----- *** Total Delta -19mV **** +7mV **** +6mV *** Percent Delta .871% .320% .275%

*Vcc = 5.065* ************** LMV431B *** LM334 **** Op Amp ** *************** TP1 **** TP2 **** TP3 *** Temperature *** Vr **** Vr **** Vr *** *************** ***** **** ***** **** ***** *** *** 72 F **** 2.176 **** 2.180 **** 2.181 *** *** 74 F **** 2.169 **** 2.181 **** 2.181 *** *** 78 F **** 2.169 **** 2.182 **** 2.180 *** *** 82 F **** 2.167 **** 2.182 **** 2.186 *** *** 86 F **** 2.164 **** 2.181 **** 2.186 *** *** 90 F **** 2.163 **** 2.180 **** 2.186 *** *** 95 F **** 2.160 **** 2.183 **** 2.186 *** *** 100F **** 2.159 **** 2.183 **** 2.187 *** *** 105F **** 2.154 **** 2.184 **** 2.186 *** *** 110F **** 2.153 **** 2.183 **** 2.186 *** *************** ----- **** ----- **** ----- *** Total Delta -23mV **** +4mV **** +6mV *** Percent Delta 1.05% .183% .275%

It is likely that some of the change in readings was due to the meter. Note the cases where the slope changes polarity for both the LM334 and the op amp sources. I can't prove it, but that looks like measurement error to me. Note also that the slope was consistently negative for the LMV431

The next test was short (mA meter connected from source to ground) vs load (2181 ohm resistor in series with mA meter).

The LMV431B circuit went from 1.003 mA (load) to 1.005 mA (short). The others did not change when the load was shorted out. This test was repeated many times, to rule out any change in the reading being caused by the meter itself.

10 turn trim pots and 1% fixed R's were used. Everything is soldered to a PCB except the test resistor and the op amp (LT1014), which is socketed.

I'm interested in improvements that could be made to the test setup and approach, and to the circuit designs for the test.

Ed

Reply to
ehsjr
Loading thread data ...

One thing I'd add is an AC test; transformer-couple some (1 kHz sinewave) AC into the positive power supply lead, and measure the AC that appears on the load resistor. This will give you an output impedance, and it can be modified to test for power supply feedthrough at higher frequencies.

Reply to
whit3rd

Interesting, rather than use a transformer (unless that _is_ easy), you could modulate an LM317 regulator feeding the supply? Inject AC into the tail of the voltage setting resistor, near ground.

You could feed a signal (square wave) from a 555 into there, same result, a perturbed supply to check on output impedance or PSRR, whichever way you like to call it.

Grant.

Reply to
Grant

This might be useful:

formatting link

Paul=20

Reply to
P E Schoen

Yeah, so is a green LED and a pnp BJT, but the OP wants to do things their own way.

Grant.

Reply to
Grant

I'm looking for improvements over what I have already for both the test circuit design and testing approach.

The LT3092 Paul suggested is a good answer for this post and it could "join" the testing. whit3rd's suggestion of injecting AC via a transformer, and your suggestion of modulating a 317 adj pin are good "appetite whetters", which I plan to try. Perhaps the green LED + PNP should join the testing as well. Do you think it will perform as well as the LM334 circuit? I've used LED's before in crude regulator circuits, but not in a circuit that has to be more precise and stable.

Ed

Reply to
ehsjr

Any circuit with a pass transistor is crude because you don't account for the base current. The (old, ordinary) green LED and transistor (npn or pnp depending which way you want the current) are supposed to cancel Vbe drift with temperature -- I have no idea how good the output impedance is -- but that circuit you put up with the '531 was very odd, I thought. So I don't know what you're really up to.

Voltage reference, opamp + FET probably the best, but make sure the opamp common mode range is there when you do a source from

+ve rail circuit.

I have an interest in making a decent current source, sink, but at the moment other things distract me, I'll get to it and report on it -- don't know when.

Grant.

Reply to
Grant

I finally got a chance to try your idea, at least in part. I started with a 60 Hz sine wave at 500 mV on the DC input. The 431 + 2 xsistor circuit produced over 250 mV garbage on the output. The LM334 circuit produced the same output whether or not the AC was on the DC input or not. The op amp circuit produced about 6 mV garbage on the output, but that circuit is powered by the DC input line that has the AC riding on it. I need to power it from a separate supply that doesn't have the AC riding on Vcc to make the op amp results valid.

Based on what I found at 60 Hz, I decided not to go to the

1 kHz test, at least until I can modify the setup so that the op amp Vcc will be clean. It's interesting that your test instantly and dramatically revealed the weakness of the 431 + transistors circuit as compared to the 334.

Ed

Reply to
ehsjr

I finally got a chance to try your idea, at least in part. I started with a 60 Hz sine wave at 500 mV on the DC input. The 431 + 2 xsistor circuit produced over 250 mV garbage on the output. The LM334 circuit produced the same output whether or not the AC was on the DC input or not. The op amp circuit produced about 6 mV garbage on the output, but that circuit is powered by the DC input line that has the AC riding on it. I need to power it from a separate supply that doesn't have the AC riding on Vcc to make the op amp results valid.

Based on what I found at 60 Hz, I decided not to go to the

1 kHz test, at least until I can modify the setup so that the op amp Vcc will be clean. It's interesting that your test instantly and dramatically revealed the weakness of the 431 + transistors circuit as compared to the 334.

Ed

Reply to
ehsjr

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.