Battery powered Current sink with low voltage cutoff

[gregfox]

Hi Kris, Well at last I finally got your design together and after fixing two wiring mistakes got it working. There are a couple of glitches;
When the switch in placed in the ON position the circuit lights up, even with no power on STK2, but when the switch is returned to AUTO it turns off. I checked the wiring and can find no mistake.
I found I could only get ~2V on the DIGITAL VOLTMETER connected to pin 2 of U1A. I think it is because of the 1K resistor RR. I believe you put RR in series with U2 to limit the current. I also noticed that as VRT is turned up, the voltage on pin 3 of U2 would vary quite a bit, from ~8v to 1.7V
If I replace RR with a 300 Ohm resistor I’m able to get the full 8 volts on the meter, however the current draw goes from 35mAH to ~300+ mAH, which would deplete the battery pretty fast.
In addition, I’m playing around with a 47K 10 turn (don’t have a 10K) pot for VRC. What formula was used to determine the topping resistor with VRM?
I have not been able to test if it will turn off at a set voltage because SW1 must remain on.
Any thoughts, I’m about fresh out?

 

[KrisBlueNZ]

So you're saying that even with plenty voltage on SKT2, the circuit turns off when SW1 is returned to the AUTO position? That's not right.

Re the reference voltage jumping around. Have you tried without the digital voltmeter connected? It sounds as if the voltmeter may be affecting the voltage. It should have a high input resistance and shouldn't affect the voltage, but try disconnecting it anyway and see whether the 8V becomes stable.

Re VRC, the requirement is for 1.5V on the top end (to give you 1.5A maximum sink current). The 33k resistor and VRM, together, must drop 6.5V and VRC must have 1.5V across it. So the ratio between these two resistances (the 33k plus VRM, and VRC) must be 6.5:1.5. Another way to calculate new values if VRC is 47k is just to multiply the other resistances by 4.7 as well, so VRM would become a 47k trimpot and the 33k resistor would become a 150k resistor.

Can you post some photos of your circuit?

 

[gregfox]

Photos

I’ve attached some photos of the circuit, I know not very neat, but it is what it is.
The answer to your question, yes with plenty voltage on SKT2, the circuit turns off when SW1 is returned to the AUTO position.

I tried disconnecting the voltmeter, to no avail. It’s impedance is 100K.

VRC responds as long as VRT is at the Min. setting, and one can change the current, but that too is dependant on the voltage at U2 pin 3

Any ideas? Should I build it up on a better board? Should I try a AP431A ADJUSTABLE PRECISION SHUNT REGULATOR (250 mA)

As I see the it right now;
A. SW1 must be ON to work
B. 8V reference all over the place when VRT moved up from min. To max.
C. Can’t test turn off because of (B) above.
Thank you for the help!

 

[KrisBlueNZ]

I'm not very happy with your construction. I think you should redesign the layout on paper first, and aim to keep the component leads short and the axial components mostly oriented the same way; that is, perpendicular to the tracks.

You may find it helpful to concentrate on components that are closely connected in specific sections of the design, and create a tidy, optimised layout for that section. Then you can place the sections on the board and connect them together.

Part of the problem with your layout is that it's impractical to verify because of all the components crossing each other and connecting between unrelated points. If you plan it all out on paper and keep those aims in mind, you should end up with a layout that is clearer to follow, and is properly documented.

If you have experience with an image editor that supports multiple layers, you may want to use that to create the layout. This will require quite a bit of initial work creating the component images, but it may work out better in the long run.

There are several programs available for stripboard layout entry but I've tried all of the free ones, and didn't find any of them very useful.

 

[gregfox]

I knew you were going to say that, so I’m already Planning the layout in COREL. I will rebuild it the way it should have been done in the first place. I will send pix when done.
BTW I did a mock up of the U2 part of the circuit, and the slightest (3-8mA) will change the voltage at U2 pin 3 from 8V to 3-5V depending. I don’t see why there should be any current draw at all. Do we have to redesign the U2 circuit?

 

[KrisBlueNZ]

I'm not sure what you're asking. Do you mean that if you draw 3~8 mA of current from the 8V reference rail, it will drop down to 3~5V? Yes, that's true. No part of the circuit is supposed to draw significant current from the 8V reference rail. The three resistive loads (the two potentiometers and the voltage divider that sets the 8V rail voltage) should draw less than 1 mA between them, and the 1k resistor from the 8V rail to the 11.1V rail should supply about 3 mA. So there shouldn't be a problem.

 

[gregfox]

I agree, there shouldn’t be a problem, but we are drawing around 4-5mA now, The TL431ACLP is good for 10mA, I was thinking of using a beefier part like the AP431 or AP431A which can do 200mA just to be safe. Anyway that was my thinking…

 

[gregfox]

An unrelated question; I have a bunch of 47k 10 turn pots, and I noticed when you crank them all the way up to 47k or so and then a bit more the meter reads OL, so there is no resistance. Is this normal or is it a defect?

 

[KrisBlueNZ]

What do you mean that we're drawing 4~5 mA now? There are three loads on the 8V reference voltage rail, and none of them draws more than 0.5 mA.

Re the pots, to my knowledge they should never go open circuit, so I would regard it as a fault. Do you have a manufacturer and part number, or a link to a data sheet? That would be the way to know for sure.

 

[gregfox]

That might of been the problem, because before I tore it down I measured 3.45 mA at the 1K resistor. I might have had a solder bridge, but too late now. I finished the paper layout and will start the new (clean) board today. The pots are "BOCHEN WXD3-13-2W"

 

[KrisBlueNZ]

You would have 3.45 mA flowing in the 1K resistor. Most of that current flows into the TL431 to allow it to regulate. A small amount of it flows into the potentiometers and the voltage-setting resistors.

I've Googled that part number and found it for sale, but I didn't find any data sheet. If you can find one, please post a link.

 

[gregfox]

Will do, if I can find one.

 

[gregfox]

Hi Kris,
I know you spent a lot of time on this project, and I appreciate your time.
I’ve built up a new board, and it too was not successful. The problem seems to be everything to the left of the dotted line doesn’t work, everything to the right works well.
I decided to bread board the attached circuit and I indicated the voltages at various points.
When power is applied the circuit turns on and SW1 has no effect at all. It doesn’t turn on or off, it just stays on. I spent the last 3 days working on this and I can’t find an answer.
Any help is greatly appreciated.
Regards,
Greg

 

[gregfox]

Some pix on new and cleaner circuit

 

[KrisBlueNZ]

OK, thanks for the diagram with voltages marked. That's very useful.

This may seem like a silly question, but are you testing with the TLE2022 installed? The voltage on TLE2022 pin 1 is neither high nor low, and if the TLE2022 is present, it should be driving that pin either high or low.

The other possibility is that it's oscillating. Do you have access to an oscilloscope? If so, look at the signal on TLE2022 pin 1.

 

[CDRIVE]

Some pix on new and cleaner circuit

Judging by the work on the component side of the board I'm sure the solder side is looks just as good but just to be sure you're not a "Bigger the glob the better the job" solderer can we see the bottom of the board? If my buddy Ron reads this he'll be ticked off at me!

The other possibility is that it's oscillating. Do you have access to an oscilloscope? If so, look at the signal on TLE2022 pin 1.

Ooo I like this thinking.

Chris

 

[gregfox]

Thanks for the reply Keis and CDRIVE,
Yes the TLE2022 was in the circuit. I do have a complete electronic workshop and when I put it on the scope there was only about 10-15 mV of ripple, probably coming from the power supply. Hooking it to a battery made no difference.
I don’t think I made any mistakes in the wiring, as I checked it 3 time then went so far as to bread board it in isolation, including replacing all components with the same results.
The old circuit I made worked (as far as on/off) but didn’t turn off well. I’m still stumped.
Regards,
G.
PS I’ll soon post the solder side, but I checked that also many times.

 

[gregfox]

Bottom pix...

 

[KrisBlueNZ]

Thanks Chris for your suggestion. I think the underside looks pretty good, don't you?

Well, 3.2V on pin 1 of the TLE2022 is definitely wrong. Did you say you looked at that pin with a scope and it was clean apart from a bit of ripple?

Could you check the voltages on pins 1, 2, 3, 4 and 8 with a multimeter, and also look at those pins with your scope. Switch the scope through the full range of timebase speeds. Sometimes oscillations won't show up well if the timebase is running too fast. Then report the voltages on all those pins, and a description of the waveform you see on them as well.

Do you have an image of how the board is supposed to be laid out? It's very hard to follow the photographs to check the layout.

Why do I see a 68k resistor in between the trimpots? There aren't any 68k resistors in my schematic. It looks like it's an ancient 10% carbon resistor, and there's another one in that style mounted on end near the TL431. You might want to check their values if they're very old.

I can't suggest anything else at this stage.

 

[CDRIVE]

Bottom pix...

Well I didn't see any giant solder globs bridging traces!

Chris