0
\$\begingroup\$

I'm trying to figure a way to switch a DPM from battery to battery in a 10 battery series string. At first I thought I'd do it manually using a small SMD 2 pole 10 position rotary switch but the only ones I've found at a reasonable price & size are coded (BCD). Then I thought it would be better to automatically scan. I bet there is a device out in chip land that could do the deed.

I would prefer to not go smart (i.e. PIC) or relays if possible. Something simple, small & cheap. Discrete FET switches would take up a fair amount of physical space which I don't have. Is there a 20 FET ASIC?

What is a circuit I could use to switch between multiple batteries? diagram

\$\endgroup\$
  • 4
    \$\begingroup\$ You want something cheap, and rotary switches are too expensive. You don't have the space for discrete FETs, and you don't want a micro. ...But you want a 20 FET ASIC? Good luck. \$\endgroup\$ – Bort Mar 28 '17 at 18:32
  • 1
    \$\begingroup\$ what's wrong with this puppy from DigiKey grayhill.com/assets/1/7/Rotary_71.pdf \$\endgroup\$ – Trevor_G Mar 28 '17 at 18:48
  • 1
    \$\begingroup\$ If you really want to automate it, as you say, then you could use a microcontroller with a 10+ input ADC. Put a potential divider across each battery's positive terminal and the lowest battery negative (call it GND) and connect each divider output to your ADC. Hang an LCD display off it, or pick a demo board with a display on, and you're in. Some software to write and some learning to do but a fun project and good to learn from. All depends on what you personally want to do. BTW, don't ask for microcontroller recommendations, there's a big wide Interweb out there. \$\endgroup\$ – TonyM Mar 28 '17 at 18:53
  • 1
    \$\begingroup\$ Since you want small and cheap, you're stuck with mechanicals. See Trevor's comment for a good start. Any multiplexing IC/ASIC is going to run up against the fact that you have 120 volts from one end of the battery string to the other. That is not something you'll find a solution to easily. \$\endgroup\$ – WhatRoughBeast Mar 28 '17 at 18:57
  • 2
    \$\begingroup\$ You need a switch with break before make operation to avoid short circuits. If you use an electronic switch, it should be rated for operation with potential differences up to 120 V. \$\endgroup\$ – Uwe Mar 28 '17 at 20:18
1
\$\begingroup\$

There are all kinds of solutions to automate this. Two spring to mind.

If your accuracy is simply .1V then you could use 10 resister voltage dividers going from the top of each battery to ground and multiplex that into a micro, measure the voltage sequentially on the top of each battery and do the math there and show the results on some suitable display.

You could use linear optical isolator across each battery to translate the voltage on each down to the right level to feed via a multiplexer or slide switch into a meter or again into a micro/display combination.

Or if you only care about battery volts <10V or some such, maybe just a led indicating that for each battery. Maybe even 10 bar-graphs of leds....

enter image description here

However, since you mention trays of batteries, and unless you want to collect the data in a computer or something, the cheapest (counting your time), quickest, and least effort solution would probably be to just buy 10 cheap volt meters.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

Three hurdles are presented here:

  1. the voltages are high (over 100 VDC), and potentially hazardous
  2. the display only gives a voltage, but that information is useless unless you also know which taps are being probed (i.e. which 12V section)
  3. the 'DPM' presumably is a low-voltage voltmeter, which requires its own logic and bias power and some kind of ground reference, and attenuation of the input voltages.

There are solid-state high voltage switches that can select one tap of the eleven in your series, see CPC7601 , and which might be powered from the HV string itself. The DPM can be replaced with a microprocessor and small display, which can give both a digitized voltage and indication of the battery section. And, there are operational amplifier circuits that can buffer the battery voltages down to manageable levels so that the microprocessor can handle the input.

It will take three or more ICs and a display, and some programming, to make this happen. It would be less expensive to label test points, and poke them with voltmeter probes, of course, but that exposes the terminals and is a shock hazard. A mechanical switch, while appealing, is going to require mechanical mounting and a suitable enclosure; it puts ones fingers near those high voltages.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

Thanks for all the input. I'm thinking my initial approach isn't practical due to availability/cost of the switch (mechanical or electrical). What I'm trying to do is connect a battery health DPM (from China - where else?) to individual 12v AGM batteries in large UPS systems. There are 8-10 batteries in an enclosed tray with 4 to 32+ trays. Currently, the health of individual batteries can't be determined online. I'm thinking of a device that could mount on the front of each tray and scan each battery. The DPM will show its health. How to scan is the problem. Below is a concept I cobbled. I realize the 74HC4067 will let its smoke out with the voltage involved. The CPC7601 is too much $. Maybe I'll be ale to fit some SOT-23 packages in the space I have to work with. enter image description here

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Use a plain old 4067 instead and you are fine. These are ok at 15V and s̶t̶a̶r̶t̶ ̶t̶o̶ ̶s̶m̶o̶k̶e̶ explode at ~18V. \$\endgroup\$ – Janka Mar 29 '17 at 20:17
  • \$\begingroup\$ @Janka.... how about at 130V DC? \$\endgroup\$ – Trevor_G Mar 29 '17 at 20:42
  • \$\begingroup\$ Yeah, one would need another switch to get the power supply from the battery in question, too. This should be very tricky. \$\endgroup\$ – Janka Mar 29 '17 at 20:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.