21
\$\begingroup\$

I've been reviewing this reverse engineered schematic of a bmax6 style lipo charger.

To my understanding this circuit is used to dissipate excess charge in a cell when balance charging. The circuit uses a parallel resistor network as the current sink (R146, R152, R158, R164, R174, R180). It appears that all 6 resistors are 120Ω, and thus the network has a total resistance of 20Ω (if I did the math right).

Balance Charging circuit showing a parallel resistor network

Why would circuit/pcb designer chose this configuration over a single 20Ω resistor? What are the pros and cons to consider when doing this?

\$\endgroup\$
3
  • 6
    \$\begingroup\$ Probably cheaper, especially if these are SMD and the single high power resistor is wire leaded. \$\endgroup\$
    – user16324
    Oct 4, 2020 at 20:49
  • 10
    \$\begingroup\$ In other applications, a common reason is to save on number of different components. Taking three 10k resistors from a single reel is machine time, while putting two reels with 10k and 5k into the machine is operator time, so if you put two 10k in parallel, you save on manual work during assembly. \$\endgroup\$
    – Simon Richter
    Oct 5, 2020 at 15:22
  • 2
    \$\begingroup\$ Optimized solutions are frequently incongruous with human intuition. It's precisely because we're bad at thinking up creative ways to be efficient that optimal solutions often look silly to us. 99% of the time, when you see something weird like this, it's because it saved someone time, money, or both. People tend to get stuck on the engineering considerations and forget about all the ways reality conspires to frustrate the ideal on-paper design. When you get used to looking for efficiency quirks, these types of questions become easier to figure out on your own. \$\endgroup\$
    – J...
    Oct 6, 2020 at 10:13

6 Answers 6

Reset to default
55
\$\begingroup\$

Power dissipation will be the driver.

  • Using six in parallel allows use of standard resistors which may be a stock item.
  • Using standard parts allows use of automatic assembly equipment.
  • Lower profile.
  • Heat spread out over larger area resulting in lower peak temperatures.
  • Ability to combine to make a non standard value. The 20 Ω in your question is not an E12 value so it's probably not available in wire-wound.
  • Reliability: if one fails the circuit might continue to operate - but a cascade failure is likely.

Thanks to my little helpers below!

\$\endgroup\$
9
  • 1
    \$\begingroup\$ E12 comes to mind too. \$\endgroup\$
    – winny
    Oct 4, 2020 at 19:53
  • 2
    \$\begingroup\$ @winny I always wondered about this, since I've only ever done design for medium volume where the cost difference is effectively 0 (<2M Units), how much cost does it actually save per resistor using E12 vs E24? \$\endgroup\$
    – BeB00
    Oct 4, 2020 at 20:22
  • 1
    \$\begingroup\$ @BeB00 You might have to take in to consideration bonuses for accounting department personnel who can demonstrate apparent cost savings. \$\endgroup\$
    – Andrew Morton
    Oct 4, 2020 at 20:26
  • 11
    \$\begingroup\$ @BeB00 When you work for a large company with many products, leadtime for odd components versus stock of existing ones in production becomes a real thing. \$\endgroup\$
    – winny
    Oct 4, 2020 at 21:20
  • 1
    \$\begingroup\$ Another factor is that a bad solder connection might reduce net conductivity by 16%, but that would be better than having it reduce net conductivity by ~100%. \$\endgroup\$
    – supercat
    Oct 5, 2020 at 16:50
24
\$\begingroup\$

Let's have a look at the pcb...

enter image description here

It looks like they used several resistors for higher power dissipation. This is quite common, as several low power resistors may be cheaper than one high power resistor, especially if you already use the value somewhere else in the design, which means you already have them loaded on the pick and place machine so you don't have to load an extra reel just for the power resistors.

Also several small resistors can be spread over a larger area, so they make less of a hot spot and get more air cooling. If there are thermal vias to copper used as a heat sink on the other side of the board, spreading the resistors around will also spread the heat on this copper heat sink.

Notice the yellow JK30 thru-hole component sitting right above the resistors. It's a PTC resettable fuse. When the resistors heat it, its trip current threshold will get lower. Perhaps it is used as a temperature sensor to prevent the resistors from overheating... but it only senses the temperature of the top two rows of resistors.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Parallel resistors are a common approach in balancers. As you say spreads heat over a larger area. \$\endgroup\$
    – StainlessSteelRat
    Oct 4, 2020 at 21:32
  • \$\begingroup\$ What package size are the resistors shown here? I see two different sizes. 805 and 1206? \$\endgroup\$
    – Mazaryk
    Oct 5, 2020 at 3:29
  • \$\begingroup\$ I liked this idea - until I saw a capacitor was bent over and covering 1/3 of the resistors. \$\endgroup\$
    – UKMonkey
    Oct 7, 2020 at 23:08
8
\$\begingroup\$

An additional point that noone has yet mentioned: Redundancy

If one high power resistor fails, the circuit is probably affected significantly.

If one of your 120 ohm resistors fails (and nothing else does) then the 20 ohm effective resistance increases to 25 ohms, not an open circuit.

Of course the chance of a single component failing and not causing general upset, well that's low but not impossible.

\$\endgroup\$
7
\$\begingroup\$

Parallel resistors, greatly imbalanced in value, can be for trimming. If 1% at 1,000 ohms, then 1MegOhm in parallel will reduce the total by 0.1%.

Parallel resistors will have a greater area, thus more vulnerable to electric field aggressor flux inflows.

Parallel resistors may have more underlying Planes, to which HEAT CAN BE DUMPED through the insulating epoxy-fiberglass substrate. FR-4 has about 200X the thermal resistance of copper, but thin sheets (1/16 inch, 1/48 inch, etc) are the distance of these PCBs.

Parallel resistors may be needed to reduce Thermal Distortion, where for Audio material (or music) the bass notes will greatly modulate the RESISTANCE and likely change the gain. This gain change will reflect onto the high tones, as AM sidebands.

Read work by Walt Jung on sizing resistors, to reduce thermal distortion that degrades Power Audio Amplifiers.

The thermal time constant of 1 cubic centimeter of silicon (clay? ceramic base of resistors?) is 114 seconds.

The thermal time constant of 1 cubic millimeter (about the size of a SMT resistor) is 100X faster at 1.14 seconds.

The thermal time constant of 100 micron cube of silicon (perhaps the size of a large resistor on the surface of an Integrated Circuit) is 100X faster at 0.0114 seconds.

\$\endgroup\$
4
\$\begingroup\$

For typical component sizes, E12 vs E96 or the selection of 5% or 1% components makes very little difference, since the cost of placement is much more significant than the cost of the resistor. However, while a 4000-part reel of resistors up to 1206 costs a few USD, prices go up very rapidly for larger components, although the performance does not. Using several small components is advantageous as it spreads heat better, but for a mass-market product the cost of manufacture is more significant.

\$\endgroup\$
4
\$\begingroup\$

As already mentioned, parallel resistors (of a given power rating) offer greater dissipation than a single one. Saves having to inventory 'special parts' too.

Another advantage is that any heat generated is spread over a wider area. Helps avoid 'hot spots' on a circuit board.

You can also use parallel or series resistors to produce unusual resistance values without resorting to purchasing specific values.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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