2
\$\begingroup\$

If we have 3 PCBs connected in series (i.e, negative terminal of one is connected to positive terminal of the next PCB), and this is connected to a 36 V DC supply, can we assume the voltage drop across each is 12 V DC?

Application: Have a solar panel with a very-high voltage DC (~200V) and high-current (~30A) and trying to see if I can get away by connecting multiple Bitcoin hash boards in series without expensive buck converters?

\$\endgroup\$
4
  • 14
    \$\begingroup\$ Unless the PCB has an extremely simple circuit on it (like a single resistor), no. \$\endgroup\$
    – vir
    Jul 25, 2023 at 21:08
  • \$\begingroup\$ This will only work for constant power devices like non-thermostat heaters and (dumb) lamps \$\endgroup\$ Jul 25, 2023 at 23:16
  • 1
    \$\begingroup\$ With multiple Bitcoin hash boards you should never try this. They are no pure resistive load with equal resistance. It is a very bad idea and you should forget it now. \$\endgroup\$
    – Uwe
    Jul 27, 2023 at 14:46
  • \$\begingroup\$ You have discovered that proper prior planning prevents poor performance! If you'd known that you wanted to power 12V equipment you might have planned your solar panel array to make approximately 12V instead of 200V. Is it too late to change it? \$\endgroup\$
    – user253751
    Jul 27, 2023 at 17:40

6 Answers 6

8
\$\begingroup\$

Assuming that when you say "connect in series" you mean "connect each board's one and only power input in series with the next ones one and only power input", then all of the above, plus:

Typically, the power supply on a PCB is referenced to the circuit's local ground and things just get weird if you offset that local ground from the rest of the world -- think plugging a USB cable into the board and having sparks fly and smoke come out.

There may be some bizarre application where you'd want to do this, but it would be very difficult to draw just the right amount of power for a board without starving the rest of the boards of either current or voltage.

\$\endgroup\$
4
  • 3
    \$\begingroup\$ One application is 4-20mA industrial control loops, which connect modules in series across a voltage supply, each draws power from the loop and communicates by changing the current in the loop (mostly analogue but there were digital ones in the late 1980s, my first job was working on Pakscan loop controllers rotork.com/en/products-and-services/control-networks/… ). \$\endgroup\$ Jul 26, 2023 at 10:53
  • \$\begingroup\$ In an undersea communication cable the optical fiber repeaters are all connected in series with one end at at + high voltage and the other at a negative. See en.m.wikipedia.org/wiki/Submarine_communications_cable \$\endgroup\$
    – D Duck
    Jul 26, 2023 at 15:48
  • \$\begingroup\$ The easiest way to make this work reliably would be to either burn some power at each repeater for better stabilization, or to drive a constant current through the cable -- or both. Doing it with a fixed voltage is really where things would get odd -- but maybe they make it work. \$\endgroup\$
    – TimWescott
    Jul 26, 2023 at 17:07
  • \$\begingroup\$ @DDuck The sea cable repeaters have been designed for serial power input connection since the very first repeaters many decades ago. Not only for optical fiber cables but also for coaxial cables since en.m.wikipedia.org/wiki/TAT-1 in 1956. \$\endgroup\$
    – Uwe
    Jul 27, 2023 at 14:44
9
\$\begingroup\$

can we assume the voltage drop across each is 12V DC?

No you can't because each PCB may appear as a different load resistor and, the PCB with the highest resistance between its power terminals will have proportionately a higher voltage across it.

\$\endgroup\$
6
  • \$\begingroup\$ Thanks - what if the PCBs are exactly the same, i.e., they're either ON/OFF and have similar power profiles? I'll edit the question for more clarification on the application as well. \$\endgroup\$
    – necro
    Jul 25, 2023 at 23:24
  • 5
    \$\begingroup\$ @necro For all we know the PCBs are nonlinear loads or even have negative resistance - as would be typical if there was a buck converter on the input. Without telling us what those PCBs are, we can't help you really. \$\endgroup\$ Jul 25, 2023 at 23:37
  • \$\begingroup\$ @Kubahasn'tforgottenMonica - (edited the question) but these are Bitcoin mining boards. A typical board has multiple ASIC's connected in series and parallel configurations so was curious if we can extend this principle at the PCB level as well. Thanks! \$\endgroup\$
    – necro
    Jul 26, 2023 at 0:35
  • 1
    \$\begingroup\$ @necro you can’t apply the same reasoning between two totally different situations. Your bitcoin miner does not connect the power of the ASICS in series. There are standard solutions for obtaining AC or DC supplies from solar panels which are most likely cheaper and more reliable than trying to home brew a solution. \$\endgroup\$
    – Kartman
    Jul 26, 2023 at 6:49
  • 5
    \$\begingroup\$ Your boards are not on or off. The consumption will vary dependent on what the ASICs are doing. \$\endgroup\$
    – vidarlo
    Jul 26, 2023 at 9:38
3
\$\begingroup\$

If you have any active device on that PCB, no you can't (active device include diodes, transistors, IC, regulator or basically anything that isn't a resistor, capacitor or inductor).

\$\endgroup\$
2
\$\begingroup\$

In addition to what others have said...

You seem to misunderstand the nature of voltage supplied to PCBs. You're treating them as if they create the voltage drop like series resistors.

They do not.

I used to design BiCMOS circuits. Parts of those circuits were heavily bipolar (in every sense of that word...). Bipolar transistors are not like field effect transistors (FET). We depended on the supplied voltage so we could step it down using diodes and VBE drops as part of the circuit.

In other words, the supply voltage was required by the circuit, not created by the circuit. For example, for those circuits that used a 5VDC supply, you could not put a 100-ohm resistor on the supply line and hook it up to a 6VDC source, then expect 5VDC to appear at the true positive input for the circuit. What was more likely to happen is 2.2VDC and a wholly inoperative circuit.

If you chain those PCBs together as you suggest with the voltage capacity of those panels, I expect them to burn up. I don't know if you need the expensive buck converters — but at the very least you need an appropriate transformer so you can hook all the PCBs together in parallel.

\$\endgroup\$
0
\$\begingroup\$

If the resistance in the PCB's is the same then yes, otherwise no. The PCB resistance will be probably close to the same, the contact resistance will vary more.

Putting the boards in series will be possible if the outputs are isolated and the boards draw equal power, otherwise no, loading variation will cause different voltage drops across the devices and could cause problems or operation beyond recommended voltage and current values.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ This is true if the boards have a well-defined resistance, but quite a lot of devices don't have a well-behaved relationship between voltage and current like resistors do. \$\endgroup\$ Jul 26, 2023 at 9:57
0
\$\begingroup\$

Many devices will need to usefully employ varying amounts of power under varying conditions. There are at least four ways a device with varying energy needs may accomplish this:

  1. A device may be designed so that it will pass as much current as necessary to extract the necessary amount of energy per second from the applied voltage, so long as the voltage is above a certain minimum below which the device will shut down.

  2. A device may be designed so that it will pass as much current as necessary to extract the necessary amount of energy from a certain nominal voltage, with degraded or unreliable operation if the voltage falls below that, and with any excess power resulting from excess applied voltage being converted to heat.

  3. A device may be designed so that it will pass a certain amount of current, and drop as much voltage as necessary to extract the required amount of power.

  4. A device may be designed so that when a voltage is applied within a certain nominal range, it will draw an amount of current that will be sufficient to satisfy its maximum energy need, and will dissipate a variable amount of unused energy as heat.

The first two kinds of devices are much more common than the last two, but only the last two may safely be wired in series as you describe. Then a number of devices are wired in series, it's not possible for any of them to pass any more current than the others. If some in a series chain reduce their voltage drop when they're receiving too little power, while other devices increase their voltage drop when they're receiving too much power, the result will be that the devices which were receiving too much energy will get more, and those that weren't getting enough will get less. If, however, devices in a series chain that are getting insufficient energy increase their voltage drop, while those that are getting too mcuh decrease theirs, and if the supply for the chain reduces output voltage when the current draw is above nominal and increase it when it's below nominal, however, things will fall into a balanced equilibrium state.

\$\endgroup\$

Your Answer

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

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