Replacement Microscope Dimmer

[JohnSL]

My father was visiting recently and gave me a project, which he'd like back in a couple of weeks. There is a doctor who spends about half of her time at a clinic in South America, and the light in one their microscopes got fried. The circuit was designed for 110V and someone apparently plugged it into a 220V outlet. They use solar power, and the output is 12V, which they then feed into an inverter.

My plan is to replace the circuit in the microscope base with an external dimmer that takes 12V in. The bulb is a 6V, 10W Halogen bulb.

At Vetco (my local store), I found a Velleman LED dimmer that seems like it would be a perfect starting point: http://www.vetco.net/catalog/product_info.php?products_id=12740&keywords=dimmer. The specs say 12-24V input, and 12V 5A output. I'd like to modify the circuit for 12V input and 6V output.

The first step was to reverse engineer the circuit, as I couldn't find a circuit diagram anywhere for this board. They used a black PCB, which made this a little harder. Fortunately, it was a one-sided PCB. I've attached schematic I created (I haven't carefully checked it, so there could be errors).

So here is my question. Which parts should I modify for 6V output? I'm guessing the Zenar and either R7 or R8. I'm pretty rusty on this type of circuit, but I'm guessing it's using PWM, and that I'll want to change the maximum duty cycle to 50%.

[DanielS]

I think there must be a few errors in there: first, your input header's polarity seems to be reversed, then you have G3's inverting input tied to G1's non-inverting input. I'm not an op-amp guru but that seems really strange to me.

At a glance, this looks more like a linear regulator to me and that zener is in the wrong location for use as a reference for the op-amp circuit. The zener is there only to protect the FET against excessive drive voltage.

[JohnSL]

I think there must be a few errors in there: first, your input header's polarity seems to be reversed, then you have G3's inverting input tied to G1's non-inverting input. I'm not an op-amp guru but that seems really strange to me.

Yes, I was missing a connection between pins 5, 2, and R1, which I've added to the updated image. I've also enclosed the photos of the two sides of the board that I've been using to reverse engineer the circuit.

As to the input, they indeed have the negative on the top and the positive on the bottom, as you can see from the photo.

Update: I just replaced the schematic with an updated version as I realized I had the polarity on the bottom connector reversed from the PCB, so I probably had errors.

[DanielS]

Check the way your in/out headers are connected. The PCB clearly shows in+ and out+ tied together. Your diagram's D1 should be connected to in/out+ and the FET is supposed to be ground-side control instead of high-side.

Your output voltage is controlled by that pot, there may be no need to modify the circuit to adjust the outout.

Since there is no charge-pump or inductor in the circuit, you are looking at either a linear or on-off regulator. In the linear case, the FET is likely to get extremely hot and in the on-off case, your 6V bulb would be getting the full 12V during on-time. Either way, I doubt this circuit would be suitable.

You may have better luck buying a $5 2.1A USB car adapter and modding the feedback resistors to bump the ratio for 6V output.

[JohnSL]

Check the way your in/out headers are connected. The PCB clearly shows in+ and out+ tied together. Your diagram's D1 should be connected to in/out+ and the FET is supposed to be ground-side control instead of high-side.

I did have the + and + connected. I've updated the circuit so the + is on top for both connectors, as this makes it easier to read the schematic. I also had some traces wrong from D1, which I've corrected.

Since there is no charge-pump or inductor in the circuit, you are looking at either a linear or on-off regulator. In the linear case, the FET is likely to get extremely hot and in the on-off case, your 6V bulb would be getting the full 12V during on-time. Either way, I doubt this circuit would be suitable.

I'm very rusty on op amps, etc. and I'm going to have to read up on it. However, with the resisters and capacitors it seems like there is an oscillator here that is creating a PWM output, and therefore is fairly efficient.

You may have better luck buying a $5 2.1A USB car adapter and modding the feedback resistors to bump the ratio for 6V output.

I'm not following how that would help. I'm looking for a dimmer, but the USB car adapter would provide a constant voltage output. Unless you're suggesting that I create a PWM dimmer from the output of the car adapter, in which case that would not drive the lamp to full brightness.

[DanielS]

I'm very rusty on op amps, etc. and I'm going to have to read up on it. However, with the resisters and capacitors it seems like there is an oscillator here that is creating a PWM output, and therefore is fairly efficient.

Yes, with the redraw, there is one op-amp wired as a comparator and that could make the circuit behave like a switcher.

You still have that issue with the whole 12V input getting dumped in a 6V lamp unless you include an LC output filter circuit to average that out.

I'm not following how that would help. I'm looking for a dimmer, but the USB car adapter would provide a constant voltage output. Unless you're suggesting that I create a PWM dimmer from the output of the car adapter, in which case that would not drive the lamp to full brightness.

I was suggesting replacing the car adapter's feedback resistors with whatever values suit the adjustment range you want. If the car adapter has a 2.5V reference and uses a pair of 10k resistors to divide the 5V output by two before feeding it to the comparator's input, you can rip out the bottom 10k resistor in that feedback circuit, replace it with a 6.8k fixed + 10k pot, and that gives you a 4-6.2V adjustment range.

[JohnSL]

I was suggesting replacing the car adapter's feedback resistors with whatever values suit the adjustment range you want. If the car adapter has a 2.5V reference and uses a pair of 10k resistors to divide the 5V output by two before feeding it to the comparator's input, you can rip out the bottom 10k resistor in that feedback circuit, replace it with a 6.8k fixed + 10k pot, and that gives you a 4-6.2V adjustment range.

Ah, I understand now--that's a great idea. If I change the USB charger to output 6V then it's really easy to create a 555 dimmer from the 6V. Very efficient, and the bulb will never get more than 6V.

Do you have a suggestion for a specific adapter where it's easy to figure out which resistor to change? I found this PDF for a reference design, and I assume that's the type of charge you have in mind: http://www.ti.com/lit/an/slva464e/slva464e.pdf

[kjs]

no, buy a ready made car charger and replace the lower resistor as said above. These are switchmode supplies and the output is then adjustable which will give you the dimming effect. No need to add a pulse-width modulator after that.

[JohnSL]

no, buy a ready made car charger and replace the lower resistor as said above. These are switchmode supplies and the output is then adjustable which will give you the dimming effect. No need to add a pulse-width modulator after that.

Thanks, I'll try that first. I was thinking that the lower limit of 2.5V would still be too bright, but I won't know until I try it.

[DanielS]

As kjs said, no need for a 555, switching the bottom resistor in the sense/feedback voltage divider for a resistor + pot already gives you whatever output voltage (or current) adjustment range you want.

In the TPS spec's reference design in figure 3, you would be looking to replace R9 with a fixed resistor that sets the maximum output division ratio and a pot for your adjustment range. It should be easy enough to locate since most single-output supplies only have one voltage divider between output and ground going back to the PWM circuit. When an opto-coupler is involved, there usually is a 431 driving it and the voltage divider of interest would be tied to the 431's reference pin.

[kjs]

no, buy a ready made car charger and replace the lower resistor as said above. These are switchmode supplies and the output is then adjustable which will give you the dimming effect. No need to add a pulse-width modulator after that.

Thanks, I'll try that first. I was thinking that the lower limit of 2.5V would still be too bright, but I won't know until I try it.

If you are lucky to get one which has a 1.25V reference........