Circuit to convert floating ground to ground reference

Thread Starter

1969merc

Joined Jul 18, 2020
2
Greetings all.

I was hoping someone might be able to help. I built the attached circuit (not my design...) to drive the tachometer movement in my vintage car. the circuit works great on bench but not in the car. I figured out why... the meter movement in the tachometer is grounded to the case, and it's not possible to 'unground'. The circuit won't work with the grounded movement.

is there a simple way to adapt the output of the circuit to be compatible with the ground referenced meter movement?

Any help is appreciated!
 

Attachments

ScottWang

Joined Aug 23, 2012
6,933
Add a pnp stage to the c of npn and the c of pnp to the meter.

It means that the output stage from npn moves to the pnp and including the diode.
 

MisterBill2

Joined Jan 23, 2018
7,032
You could abandon that circuit and use an LM2917 tachometer IC that has one side of the meter tied to common negative. The application circuit is far simpler and very linear and has stable calibration. Plus it includes a voltage regulator.
 

Irving

Joined Jan 30, 2016
1,018
Add a pnp stage to the c of npn and the c of pnp to the meter.

It means that the output stage from npn moves to the pnp and including the diode.
Good call, but I think it needs a little more tweaking to move the calibration stuff as the diode/calibration pot is in the wrong place now. Also putting 5v across pnp's b-e isn't going to help reliability.

My take on this is:

From +5v put a 10k resistor and a 2k2 resistor in series to c of NPN.
Add PNP , e to +5, b to junction of 10k and 2k2.
Move variable resistor to c of PNP, other end to meter. Move diode across meter pointing away from ground. Shunt across meter as required.
PNP is complementary version of BC547, eg BC557 or any similar

I would draw this but can't easily on phone. Suggest you do & feedback to check...

I would recommend you get circuit board as close to meter as possible and ensure a good solid ground to the meter ground point if possible.
 
Last edited:

ScottWang

Joined Aug 23, 2012
6,933
Good call, but I think it needs a little more tweaking to move the calibration stuff as the diode/calibration pot is in the wrong place now. Also putting 5v across pnp's b-e isn't going to help reliability.

My take on this is:

From +5v put a 10k resistor and a 2k2 resistor in series to c of NPN.
Add PNP , e to +5, b to junction of 10k and 2k2.
Move variable resistor to c of PNP, other end to meter. Move diode across meter pointing away from ground. Shunt across meter as required.
PNP is complementary version of BC547, eg BC557 or any similar

I would draw this but can't easily on phone. Suggest you do & feedback to check...

I would recommend you get circuit board as close to meter as possible and ensure a good solid ground to the meter ground point if possible.
Thanks.
What I described is not enough, because it just a hint for the people who has enough knowledge to figure out what I said, and it is a quite nornal and practical circuit, the npn and pnp driver as the right top below.
 

Irving

Joined Jan 30, 2016
1,018
Thanks.
What I described is not enough, because it just a hint for the people who has enough knowledge to figure out what I said, and it is a quite nornal and practical circuit, the npn and pnp driver as the right top below.
That works too. :)

Not sure if OP has enough knowledge or not. Doesn't hurt, IMHO, to give a more complete answer. I find a lot of people come here with a specific issue, and while they are knowledgeable about that and perfectly competent to build stuff to meet it, just need a helping hand with the detail.
 

Irving

Joined Jan 30, 2016
1,018
You could abandon that circuit and use an LM2917 tachometer IC that has one side of the meter tied to common negative. The application circuit is far simpler and very linear and has stable calibration. Plus it includes a voltage regulator.
That's a neat chip I'd not come across before, thanks.
 

MrAl

Joined Jun 17, 2014
7,811
Greetings all.

I was hoping someone might be able to help. I built the attached circuit (not my design...) to drive the tachometer movement in my vintage car. the circuit works great on bench but not in the car. I figured out why... the meter movement in the tachometer is grounded to the case, and it's not possible to 'unground'. The circuit won't work with the grounded movement.

is there a simple way to adapt the output of the circuit to be compatible with the ground referenced meter movement?

Any help is appreciated!
How much current does the "movement" require?
If low current then a single op amp could do it.
 

Irving

Joined Jan 30, 2016
1,018
Whilst I recognise that the single chip solution has many benefits, we shouldn't lose sight of the OPs orginal request - what can be done with the solution he has already invested time & money in? A PNP transistor and 2 resistors is still cheaper than the 8 or so parts needed to build the chip solution.
 

MisterBill2

Joined Jan 23, 2018
7,032
Going back to the original circuit posted, the emitter of Q1 is tied to ground. So if the diode D2 were moved to the emitter side of Q1, between the R4 junction and ground, with the 150 ohm series resistor, then probably the meter could be driven by the circuit working as an emitter follower. That could be a rather small change physically, and it should do what is needed.
The benefit is that whatever non-linear operation is present, presuming that the circuit is accurate with the present circuit on the bench, would be maintained.
But if the circuit is not the one that the tachometer meter part had driving it originally, there may be accuracy problems anyway.
The layout of the drawing made it a bit difficult to trace that ground connection.
 
Top