# AC Voltmeter - Consolidating Tap points using one circuit

#### phippstech

Joined Jan 27, 2020
19
Hi,

I have a AC voltmeter circuit I'm trying to use for many tap points. At the moment, I have the circuit for each tap point:

Instead of having these circuit for each tap, I'm trying to design something that will switch between tap points that feeds into one circuit:

Is there any black box modules out there that can me control by a microcontroller or should I use a Triac based system and switch those from the microcontroller

#### phippstech

Joined Jan 27, 2020
19
Following up on this, are there any ideas?

#### KeithWalker

Joined Jul 10, 2017
3,034
The simplest solution would be to use a SPST relay for each tap point (whatever that is!).
With a resistive divider, 15Mohm/100Kohm (as shown above) connected between each tap point and common, the top of the diode bridge can be connected to each voltage divider junction in turn using low voltage, low current, single pole, normally open relay. Banks of SPST low voltage relays are available in several different sizes, specifically for use with 5V microcontrollers.

NOTE: The resistive dividers have very high value resistors so they should be mounted on very clean fiberglass boards with covers to protect them from dirt and humidity. They should not be handled during assembly.

UneMJZVf&terminal_id=1843666702464a48973e368534c8eb21&afSmartRedirect=y

Last edited:

#### Ian0

Joined Aug 7, 2020
9,535
What level of accuracy are you expecting? Won't it be ruined by not knowing the diode voltage exactly?
Is the reduction in accuracy due to the diodes any worse than just using a single tap set to the highest voltage you intend to measure?

#### Jerry-Hat-Trick

Joined Aug 31, 2022
498
How about a 1 pole rotary switch? For example the Lorlin CK1049 1 pole 12 position.

If you can live with a single diode and resistive potential divider for each tap point, which I’d favour to make the control low voltage, then a processor like the Seeeduino XIAO would give you 11 analogue inputs which could be displayed on a computer screen together. The diode so you measure only half the wave but if you assume it’s symmetrical you have 12 bit resolution and fast enough to give a true RMS value.

#### Ian0

Joined Aug 7, 2020
9,535
I am also concerned about “common” and “ground”. Where are these connected to in real life?

#### ericgibbs

Joined Jan 29, 2010
18,673
hi phipp,
A uni-selector is one option, ref this link.

Using the spare contact banks with an encoded wiring set, the MCU could drive the selector via a suitable driver, also auto-selection of the input source is possible.
E

https://en.wikipedia.org/wiki/Stepping_switch

#### phippstech

Joined Jan 27, 2020
19
hi phipp,
A uni-selector is one option, ref this link.

Using the spare contact banks with an encoded wiring set, the MCU could drive the selector via a suitable driver, also auto-selection of the input source is possible.
E

https://en.wikipedia.org/wiki/Stepping_switch
View attachment 278903
This might be physically too large for this application I'm afraid. Very interesting though.

I am also concerned about “common” and “ground”. Where are these connected to in real life?
Common is the AC common line and Gnd is the DC ground on the PCB. I'm still tending to the accuarcy of the voltage meter, but mainly concerned with finding this switching circuit that will cut down the number of this ciruits on the board.

How about a 1 pole rotary switch? For example the Lorlin CK1049 1 pole 12 position.

If you can live with a single diode and resistive potential divider for each tap point, which I’d favour to make the control low voltage, then a processor like the Seeeduino XIAO would give you 11 analogue inputs which could be displayed on a computer screen together. The diode so you measure only half the wave but if you assume it’s symmetrical you have 12 bit resolution and fast enough to give a true RMS value.
This needs to be controlled from MCU we have on the design and we want to cut down on the number of analog inputs going into the current MCU

The simplest solution would be to use a SPST relay for each tap point (whatever that is!).
With a resistive divider, 15Mohm/100Kohm (as shown above) connected between each tap point and common, the top of the diode bridge can be connected to each voltage divider junction in turn using low voltage, low current, single pole, normally open relay. Banks of SPST low voltage relays are available in several different sizes, specifically for use with 5V microcontrollers.

NOTE: The resistive dividers have very high value resistors so they should be mounted on very clean fiberglass boards with covers to protect them from dirt and humidity. They should not be handled during assembly.

View attachment 278882
UneMJZVf&terminal_id=1843666702464a48973e368534c8eb21&afSmartRedirect=y
This was actually my first though but we were hoping to cut down on the board space this circuits take up, as well as cut down on the number of analog inputs that are used. We have other sensors on the board (around 40), so trying to stay away from adding MCUs.
Have any thoughts on using a triac instead of a relay?

#### KeithWalker

Joined Jul 10, 2017
3,034
You could use miniature reed relays.

#### Ian0

Joined Aug 7, 2020
9,535
A few fundamental questions that have not been answered:
1. And most important - how are you going to deal with the error resulting from the diode voltage drop?
2. what is the range of taps,
3. how many taps.

#### Jerry-Hat-Trick

Joined Aug 31, 2022
498
This needs to be controlled from MCU we have on the design and we want to cut down on the number of analog inputs going into the current MCU
How about several MCP3008 ICs - 10 bit 8 channel serial ADC. Each one would need an individual chip select pin or a 4017 decade counter if you need up to 80 tap points. Clock and data lines common to all. I'd still suggest each channel has a single diode and a resistive divider so you measure half the sine wave (avoiding the complexity of a virtual ground, although that's not difficult) and measurements should be fast enough to get a reasonably accurate RMS reading.

#### phippstech

Joined Jan 27, 2020
19
A few fundamental questions that have not been answered:
I totally agree that the voltmeter circuit needs refinement but for this post, I want to focus on what was asked. I might create another post for problems with the actual circuit.I'm expect 25VAC and 100VAC and the number of taps will vary. I have several circuits I'm breadboarding. If I have issues with accuracy with the voltmeter circuit, I will post in a separate post.

How about several MCP3008 ICs - 10 bit 8 channel serial ADC.
I want to aviod the repeat circuit consisenting of the voltage divider, diodes, etc.

So as a reminder, I want to cut down on the number of voltmeter circuits, I want the switching circuit to be control via MCU and want to circuit to take up less board space. So to @KeithWalker point, I feel the relay system is the way to go and I offered the TRIAC idea, lets meet in the middle and use a solid state relay. So I'm breadboarding something like this at the moment:

Where the GPIO expander is controled by the MCU with I2C and the only one or two analog inputs coming in.

#### Ian0

Joined Aug 7, 2020
9,535
I totally agree that the voltmeter circuit needs refinement but for this post, I want to focus on what was asked. I might create another post for problems with the actual circuit.I'm expect 25VAC and 100VAC and the number of taps will vary. I have several circuits I'm breadboarding. If I have issues with accuracy with the voltmeter circuit, I will post in a separate post.
If you refine the voltmeter, you won't need any taps.
How many bits is your A-D?

#### KeithWalker

Joined Jul 10, 2017
3,034
You would get much better long-term accuracy and stability with lower value resistors in the voltage dividers, e.g. 1.5Mohm / 10Kohm or even lower.

#### Ian0

Joined Aug 7, 2020
9,535
lets meet in the middle and use a solid state relay.
1. The leakage current may be more than your divider current
2. It may have a built-in snubber, in which case there would be so much leakage current you might as well replace it by a short circuit.

#### phippstech

Joined Jan 27, 2020
19
If you refine the voltmeter, you won't need any taps.
I have at least 10 voltage sources I have to read in each system. There is no taking away taps.

#### Ian0

Joined Aug 7, 2020
9,535
I have at least 10 voltage sources I have to read in each system. There is no taking away taps.
So you don't need a tap selector, you need an input selector?
You could use a 4067, or, better still, get an A/D with 12 inputs.

#### Gen 46

Joined Jun 20, 2021
3
hi phipp,
A uni-selector is one option, ref this link.

Using the spare contact banks with an encoded wiring set, the MCU could drive the selector via a suitable driver, also auto-selection of the input source is possible.
E

https://en.wikipedia.org/wiki/Stepping_switch
View attachment 278903
Having worked in a Strowger telephone exchange in the 1970s, I never wanted to see or hear these U/Ss again!