Battery powered Current sink with low voltage cutoff

[gregfox]

Hello all, I’m working on a project, a Current sink with low voltage cutoff. I want to be able to connect a dummy load to a project, or run tests on batteries etc. and be able to set a low voltage cut off point. The Sink will go up to 1.5A.
In the attached schematic the current sink works along with the TLP222A which powers the LM324 (2mA) when the switch is turned to ON-METER-OFF (to conserve power)
In the lower right hand corner I placed a low voltage cutoff with hysteresis. The hysteresis is for battery testing, so that when the set point is reached, the circuit will not “chatter”, when the voltage drops to the set point. The circuit works up to the output of pin 8 (IC1C). The hysteresis can be controlled by varying resistor R7.
My problem is I can’t figure out how to incorporate a panel meter that I could set (with a multi turn pot) to set the lower voltage point. I also don’t know how to turn off Q1 at the low voltage point.
R6 does allow you to set a low voltage point, but that’s not exactly what I want. Everything else works, I tested the unit to 1.5A and the hysteresis works from the set point up to about .5 Volt. You can disregard the IC2D as I was just fooling around with ideas.
Any help will be much appreciated.

 

[KrisBlueNZ]

Hi Greg,

I'm interested in this project and I want to help.

BTW thanks for posting such a thorough and clear description of what you want to do. You are one of the very few!

I designed something similar a while ago but I want to start again and design something with better performance and features.

I've looked at your design and I think there are several things that can be improved. I would like to start from scratch. Are you interested?

FYI here are the problems that I see with your design, and questions I have.

I don't see any need for the circuit to have a power switch, as long as the shutdown current is low enough. By "low enough" I'm thinking 1 mA or less. I think the circuit should have a "start" pushbutton that starts the current sink, and when the target voltage is reached, the circuit turns off until the "start" button is pressed again. This complete cutoff arrangement means hysteresis isn't needed. What do you think?

Re setting the cutoff voltage. Does it really need to be fully variable? It would be a lot easier to have a multi-position switch (e.g. rotary switch) that selects one of several preset cutoff voltages. These would be set in advance during circuit setup and testing. What do you think?

 

[gregfox]

Hi Kris,
Thank you for your kind reply. I really don’t mind starting from scratch, but let me address your questions one at a time.
A. Power switch: There is a need to have a switch as the panel meters draw about 7-10 mA each. The meters are the cheap kind you can get from eBay, and aren’t very accurate, but close enough.

B. The idea of a start button that starts the current sink and turns off when the target is reached is both intriguing, and brilliant. It would indeed remove the need for any hysteresis, although I did solve that problem, by changing R7 to 3K.

C. Regarding the cut off voltage, I would really prefer to keep it variable; as there are many times one would like to “tweak” a sink. Besides, I think it would be easier to build with a multi turn pot, then a rotary switch.

I think I might have uploaded an old schematic and will upload the latest. I would love to hear more on this, and if you have any drawings, would love to see them. I don’t work with analog much and do better in digital. Off the top of my head, a variable reference source feeding a Presentable divide-by-n counter, or something along those lines might work very well. I’ll spend time thinking about a new design, let’s try and work this out.
Regards,
Greg

 

[KrisBlueNZ]

I was assuming that the whole circuit would be powered from the battery that you're discharging (or whatever it is). The meters would need to run while the circuit is active, but when the target voltage is reached, they could turn off, right? The problem is the power supply voltage they need. They probably won't work over a range of 4~12V, will they? Are there any that will work at 4V or less? It's not hard to add a regulator to supply them.

Also can you post links to any of these digital meters that you've already bought, or are considering buying? I would like to know about them in detail.

Thinking more about the power supply situation, it certainly makes things simpler if the load has its own power source, because the circuitry doesn't need to tolerate such a wide range of supply voltage. It would be easy to use a relay or MOSFET to control power to the whole circuit, so once the target voltage is reached, the whole circuit could be powered off, and the load on its internal battery would drop to just microamps. How does that sound?

Re setting the target voltage, are you talking about a multi-turn potentiometer, or a multi-turn trimpot? A multi-turn potentiometer will be difficult to deal with, I think, because you can't see its position just by looking at it.
Edit: Duh, that's why you want the digital meter. OK, a continuously variable threshold voltage is no problem, especially if the circuit has its own battery.

 

[gregfox]

You’re right, by having the sink circuit powered, will disregard and voltage swings on the target being tested. I tested the sink section and it works great, From 1.8AMP down to “0” mA. The 11.1V LioN batteries supply enough push to drive the Mosfet well beyond it’s 1.5A limit. You will need a 30W or better heat-sink on the Mosfet.

You’re right again, the Meter will show the Target cutoff voltage, and the other meter will show the source being sinked in mA. Note the “sinked” meter will read across the 1 ohm 1% resistor (R4) to show the current being “sinked” If you choose an ammeter, instead of a milliamp volt meter R4 could be eliminated. Also note I bring out the individual cells of the battery so that I could use a balanced charger for long battery life.

As far as using a mosfet to turn power on and off, I attached a circuit I used that works great. You got me thinking, now working on a re-design, and will upload when ready, haven’t a clue where to start, but adding some digital to the circuit wouldn’t hurt. Here are the links to some of the meters I buy. See below:
http://www.ebay.com/itm/310601671625?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

http://www.ebay.com/itm/390553601519?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

http://www.ebay.com/itm/330792995101?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

http://www.ebay.com/itm/DC-Digital-...528?pt=LH_DefaultDomain_0&hash=item5aef496a00

The parameters I like are;
Battery powered.
Two meters (or one switched) for Current and voltage.
Max sink 1.5 Amp.
Selectable Voltage cut off.
Selectable Current (up to 1,5 Amp)
What think you?
NEW-> I designed a new circuit that turns on with a single pushbutton, and when the set current is reached, the whole device turns off.

 

[gregfox]

Forgot to mention the batteries I would like to use are Li-On 3.7V 14500 700-1100 mAH (AA) size.

 

[KrisBlueNZ]

Hi Greg, I haven't forgotten you! I'm still working on ideas.

I don't know what you mean about using a presettable divide-by-n counter in an analogue circuit, but I agree with the specifications in your post #5.

I really wanted to arrange the circuit so it could be powered from the voltage source that's being loaded. Using an external power source but also requiring internal batteries just seems like a waste, if there's any way to avoid it. I think that would be possible apart from your requirement to be able to set the cutoff voltage using a meter. So I'm now assuming the internal battery will be required, unfortunately.

Is there any reason why you want to use a MOSFET as the active device in the current path? I would use a Darlington transistor in this position, because I think of MOSFETs as more suitable for switching applications. The one you specified in your PDF posting IS specified for significant power dissipation, but most MOSFETs aren't. The Darlingtons I'm suggesting are BD677 (TO-126) or BDX53 (TO-220). The TO-126 device's specifications are fine, but it's pretty small; you might prefer the TO-220 device because it's physically more solid and there are probably more heatsinks designed for TO-220 packages than for TO-126.

I will post something when I have a coherent design. In the meantime, would you please look at the advertisement for the current meter at http://www.ebay.com/itm/DC-Digital-...528?pt=LH_DefaultDomain_0&hash=item5aef496a00 and see what you think about the second connection drawing. I think the polarities of the supply and the load are shown wrong, at least.

Also, for this meter, do you know how the high-current path is related to the power supply rails? I was thinking that the thick black wire would be connected to the thin black wire internally in the meter, so the thick path could be connected in series with the current shunt resistor that is returned to the negative rail of the circuit. But if the meter is arranged differently, that won't work.

This is one reason why I don't like using products like this from eBay. Generally they're poorly documented (this one doesn't even state a part number, let alone a manufacturer), and they're unlikely to be available in the long term or even medium term. You get what you pay for!

 

[gregfox]

New update to design

HI Kris,
Please forget about the divide by n, It was something I was thinking about using a complicated scheme that would be expensive and hard to implement.

There is a way to power the device from its target, but not worth the trouble. I really like LiOn or LiFePo4 powered devices because of the high capacity of the batteries. The charging is fast and this device should run a long, long time on a single charge.

I like your idea of using the Darlington, and have ordered a few BDX53 to test in the circuit. I’ll let you know when they arrive.

I’m not sure about that meters connection you mention, and will look into it further.

Other news, I took you advice, and designed a unit that does everything I (and I think you) want. I put it together last night, and it seemed to work OK. Now don’t go crazy on me, it’s only a preliminary circuit, and not tested well. It does turn its self off when it reaches the set cutoff voltage, and it has fewer switches. Only a single push button for on and off. I think this is only a base to start from, and know it could be refined. I work with the Parallax Propeller and would love to move the whole project to a microcontroller environment, and maybe one day will. Right now I’m not very good at spin, and I’m thinking of going to C as a programming language, but that’s another story.
I look forward to reading your comments and refining the design.
Regards,
Greg

 

[KrisBlueNZ]

Hi Greg,

Thanks for that diagram. I mostly understand what you're doing there but I'm confused about a few things. For a start, are the batteries connected with the wrong polarity? And is one side of them supposed to be connected to GND?

Perhaps you could write up a circuit description for me so I can understand your design better.

Here's my attempt so far.

Here's a circuit description.

The bank of three 3.7V Li-Ion cells can be charged from a balanced charger via CN1. The power supply to be loaded, or the battery to be discharged, is connected to CN2.

QP switches power to the whole circuit. It is turned on while SW1 is held in, and when QH is biased ON. The 33k RK (there are two RKs on this version of the design - oops!) turns QP off when there is no bias source. When QP turns OFF, the +11V rail to the rest of the circuit will drop away quickly.

When SW1 is initially pressed, the circuit powers up. Both digital meters start up. U2 provides a +8V rail which is used by the load circuit and the voltage threshold circuit.

VRT provides a variable voltage that determines the shutoff threshold. VRM should be adjusted so this voltage range is from about 4V to 8V. This voltage is indicated on the digital voltmeter and feeds U1A, which is used as a comparator.

The other input to U1A is fed from the CN2 voltage via RP, with extra interference from QK, which is intended to provide a reliable shutdown for the whole circuit. This is how I intend that it should work.

Initially CK will be discharged, and the QK base voltage will be 11V. After a while, current through RK will pull this voltage down to the CN2 voltage. In either case, QK and DK will have no effect, and U1A will compare the CN2 voltage against the shutdown threshold voltage set by VRT. While the CN2 voltage is higher, U1A's output remains high and QH remains ON, holding QP ON and keeping the circuit running.

When the CN2 voltage drops below the shutdown voltage, U1A's output will go low. QH will turn OFF, QP will turn OFF, and the +11V rail will start to drop. The bottom end of CK will fall in unison, pulling QK's base voltage downwards.

Once the +11V rail has dropped by about 1.2V, QK and DK will conduct, and will pull U1A's non-inverting input voltage down rapidly as the +11V rail falls. This should ensure that U1A's output remains low as the +11V rail falls, regardless of how the reference voltage and the CN2 voltage behave as the circuit shuts down and the constant current load stops drawing current.

DN stops CK's negative end from going far below 0V and ensures that CK is discharged ready for the next run. DV ensures that U1A's non-inverting input doesn't exceed its supply voltage by much, even if the CN2 voltage is higher. RP limits the current here.

This shutdown circuit is untested. I'm not sure that it will work reliably. Perhaps a better approach would be to put a zener between U1A's output and QH's gate, with a pulldown resistor on the gate, so that as soon as the 11V rail drops below say 8V or 6V, U1A will not be able to turn QH on even if its output goes high. There might be other improvements to be made here. The important thing is that the circuit mustn't start up again during shutdown because of comparator input voltage changes or comparator misbehaviour.

VRC provides a voltage adjustable from 0V to 1.5V (with fine adjustment of the maximum voltage via VRM). U1B and QS form a linear regulator that keeps QS's emitter voltage equal to the voltage from VRC; RS converts this voltage into a corresponding current.

The digital ammeter monitors current through RS directly. It contains an internal shunt resistor but its value should be much less than RS, and the error it introduces can be trimmed out using VRM.

DP protects the current sink from reverse polarity.

An LED could be added across the 11V rail but I think the digital meters will provide an adequate indication that the circuit is running.

I chose the TLE2022 because I've used it before and always had good results with it. The A suffix has a maximum input offset voltage of 0.15 mV at room temperature; much better than the LM324. It also has a low current consumption, but this is swamped by the current drawn by U2, which requires at least 1 mA for proper operation.

U2 doesn't need a decoupling capacitor and in this application I don't think there's any need for one.

What do you think?

 

[CDRIVE]

Kris, you never cease to amaze me how fast you design your circuits. I'm beginning to think that your wife and kids don't see much of you!

Chris

 

[KrisBlueNZ]

Thanks Chris
Actually that one's been in the slow cooker for a few days and I'm still not happy with it at all. And I don't have kids!

 

[gregfox]

Hello Kris, I will study your design, and will breadboard to see if it fly’s.
Yes you're right about the battery, just a slight oversight. I have included a new diagram. Thanks for your reply.

Power is turned on by S1, which energizes T1A &T1B allowing current to flow to the device. Diode D1 prevents the IRF7319 from shutting down due to the capacitance in the circuit. A cutoff voltage is set by R10, and a source PSU/Battery/any current source is connected. Current is set by R5. IC2A is a current follower feeding a divider network R2/R3 IC2B receives feedback from Q1 When the voltage of IC1D (pin 12/13) are equal IC1D output goes to high, and allows Q2 to conduct, grounding T1A thus shutting down all power.

I have it working on the breadboard, but I will study your iteration and learn what I can. Thank you for your help, and inspiration, let’s keep this going until we have it right!
Regards,
Greg

 

[KrisBlueNZ]

OK, I get the idea. I'm not clear on a few things though.
1. I don't see how D1 "prevents the IRF7319 from shutting down due to the capacitance in the circuit"
2. Q2 seems to have its source and drain reversed and there's no need for R14.
3. D3 is either backwards or connected to the wrong rail.
4. R2's wiper is not connected and there's no need for IC2A and R3.
5. There's no regulation of the control voltages so they will vary with the battery voltage.
6. C1 shouldn't be there; it will just slow down the feedback loop for the current sink.
7. There seems to be a wire drawn through the top cell of the battery.

Also you might want to take the current monitoring point from the non-inverting input of IC2B so the current can be set before the external source is connected.

Apart from that, my concerns about the circuit are that there's not enough protection for the IC inputs, but mainly that it may not shut down reliably. You need to make sure that once the shutdown threshold is reached and the decision is made to shut down, nothing can happen as the power rail disappears that could cause it to start up again.

That is the reason for the complicated and messy circuitry in my first design. I will try to revisit it and simplify the design in that area.

 

[KrisBlueNZ]

Here's an improved design.

The changes from the previous design relate to the ON/OFF control and low-voltage shutdown. Here's a description of the new circuitry.

SW1 is a three-position switch. A centre-OFF switch can be used, because the middle position, "AUTO", doesn't need a connection. The switch can be stable in all three positions, or the OFF position can be momentary (spring-loaded); it doesn't matter. The ON position could also be spring-loaded but it would be useful to have a stable position so the low-voltage shutdown can be disabled if required.

When SW1 is set to OFF, QH's bias is removed and the circuit shuts down. When SW1 is set to ON, QP's gate is pulled near ground by DF and the circuit starts up.

U1A compares the CN2 voltage against the low-voltage shutdown threshold voltage. When the CN2 voltage is lower, its output goes low, and this low level is passed through QK and DK and pulls the non-inverting input low, latching the comparator into this state, provided that QK's gate is high. It is pulled high by RK and DL, but not if SW1 is currently in the ON position. This allows the circuit to start up - while SW1 is ON, QK is disabled and U1A's non-inverting input is free to follow the CN2 voltage, so U1A's output is determined only by the voltage and is not affected by the latching action involving QK and DK.

DL is needed because without it, when the circuit tries to shut down, RK will act as a pulldown resistor on QP's gate and will prevent QP from turning off properly. DF is needed so that QK is controlled only by SW1 without interference from QH. DR protects QK's gate from possible high positive voltages; its cathode is connected to the battery positive rather than the +11V rail for the same reason that DL is required.

Edit: Added DR, corrected date for .002 revision. Schematic is now .003 revision.

 

[gregfox]

Looks really good

Looks really good, have you tried it out yet? I’ve been stuck on a bug, I’ve been working on for hours, and haven’t gotten into you design, but will soon. I’ve ordered some parts because I didn’t have all the parts you specified, and to do it justice, I think it best to use what you specked out instead of substitutes. Thank you for all you help, If the parts get here I’ll breadboard it, and give it a go.
Great project Thanks again!

 

[KrisBlueNZ]

No, I haven't prototyped it, and I don't plan to.
You're welcome Good luck! Hope to hear from you soon.

P.S. I forgot to mention. The 3-position switch part number on the schematic is the cheapest 3-position switch I could find on Digikey. The manufacturer is "E-Switch" or something. But any 3-position switch will do.

 

[gregfox]

Hi, I built my design up on a perf board and threw it in a box. So far I works.
I have your design mostly made up, just waiting for parts, can’t wait to see if it work.
Thank you again and I’ll let you know the result.
On another note I built a battery powered power supply, and want to improve it. Interested?
Regards,
Greg

 

[KrisBlueNZ]

Sure, of course! Post the design on a new thread and PM me with the URL of the thread.

 

[gregfox]

Will do soon.
G.

 

[gregfox]

TL431ACLP

Hi Kris,
I should receive my parts soon, but on studying your Schematic I found a question I would like to ask you;
On the TL431ACLP I noticed a 1K resistor (to +11) 33K in line with a VRM10K. I don't see the purpose of these units. I looked at the data sheet and was scratching my head. Would you be so kind as to enlighten me?

I also thought it might be better to bring the voltage up to 9V since the cutoff of the batteries are 9.9V What do you think?