A little help with gauge reversal circuit please.

Hi folks. Hope you can help.
I have an antique 1968 gauge and have replaced the level sender with new which of course is incompatible as the originals are no longer made.

I know this has been discussed before - http://forum.allaboutcircuits.com/showthread.php?t=29644 (many thanks to Jack-K & SgtWookie)

All was going well using an LM2904 but as I needed more current (about 100mA) I added an output PNP, but this has led to an unwanted offset of 0.8V which means I cannot get close to zero. I need to get to 0.5v and am at present on about 1.3V minimum.

Is there a way to increase current handling from the LM2904 without adding an offset or do I need another op amp?

I'm fairly new to this level of electronics and am self taught but enjoying it, so any help would be appreciated.

Many thanks.

 

1.4V on the output is probably about right assuming you're using the exact same schematic with a pnp on the low side on a single power rail.

Try reversing the OP amp inputs and using an npn instead. Although, an n-channel fet may be required in order to turn off the transistor

 

I think I've finally found a solution & have the output down to 1v which is acceptable to the gauge.
The op-amp output goes straight to the gauge as per the original design, but I've added an NPN from this junction to ground. This way it takes most of the loading. The base is fed from the -ve pin of the op-amp, and from there I've added a 100Ω resistor between it and ground.

This design is taken from half of a totem pole driver used to drive servo motors from an op-amp.

Worked out on LT spice the op amp is only giving less than 10mA, 8mA of which goes through the 100Ω resistor to supply the NPN base. Seems a quite elegant design.

I've also adjusted the resistors supplying the sender so that m= -1 ,so this does away with the scaling resistors at the input. So instead of y=mx +b it is just y= b - x . Much simpler. Using 4x 1.3k resistors in parallel I can get away with standard 1/8W safely with no heating. A bit overkill but it gives the right result with standard resistors.
A maximum of 29mA flows through the sender - much less than the original 120mA so I'm not cooking anything.

If I could find a way of drawing the design on here I would. If you can describe anything better then please do.
Thanks.

 

If I could find a way of drawing the design on here I would.

If you don't have a suitable graphics or schematic-capture program you can always draw a circuit by hand and post a photo here (using 'Go advanced'/ 'Manage attachments').