T&L Publications

I am building some adjustable power supplies for my students and have run into a problem with the digital panel meters Im using. After I received them I discovered that they cannot use a common ground to both power the unit (7-12V) and measure the power supply adjustable output (LM317). I then tried to isolate the grounds by using a small audio 1:1 transformer from the power supply transformer to a LM7809 and cap filter to power the meter. This did not seem to work. I did hook up the meter to my own benchtop power supply to the 12V output to power the meter and the adjustable output on the measured input of the meter and it worked fine. So the meter will work if the grounds are properly isolated. Does anyone know why this is a problem and how I could power the meter and measure its adjustable output without adding a battery to power the meter?

Its normal. It has to do with the op amps inside. Im not familar with any cures?

It doesn't make sense that using the 1:1 isolation transformer wouldn't work as it would be the same as using a battery. You are keeping the common on the secondary supply circuit totally isolated from the meter input, right?<p>Dean

tetanus, How about some more info on the
"digital panel meters"
that you are using. - make, part #, etc.

yes,the transformer should have done it. There is no physical link between the grounds. What I don't understand is that my commercial benchtop power supply is able to power the meter and read its own output on another channel. The meter also works fine if powered by a 9v battery and meters my homebuilt PS output. With the isolation trans., rectifier,&cap the power to it works fine. When I add the positive output from the PS without the ground it works fine (although the 1.2V output reads -6.8 on the meter due to the floating ground). When I add the ground, the output on the meter shows a 1,or the default error message on the screen. Is it possible that capacitance from the transformer is affecting the meter? How can one further isolate these grounds?

This is a C + C digital panel meter. 7-12 volt supply. I have changed the surface mount voltage dividers to read in the 0-200V scale. There is not much info out there on these things (which is how I got into this mess). This site has most of the info (almost none) http://shop.store.yahoo.com/webtronics/ ... ddigp.html
While visiting other dIgital Panel meter sight, the use of a dc/dc converter to provide isolation for the supply power came up briefly but no mention of details (such as a recommendation).web page

Such problems with their verbal descriptions leave me with a dull headache. Seems like it's time to fire up the o'scope wih leads hooked up to measure differential voltage and probe around - but where, I would not know.
(edited for my poor spelling)<p>[ January 24, 2005: Message edited by: jimandy ]</p>

Idea 1. Too obvious maybe, you didn't mention a rectifier between 1:1 transformer and LM7809.<p>Idea 2. More likely the small audio transformer is too small. It's impedence is to high to pass the current needed by the meter.<p>What are current needed by meter, impedance and DC resistance of transformer windings, and power rating of transformer?

the meter only needs 1mA and powers up fine with the 1:1 trans and the 7809. I don't know what the impedance and power rating for the trans are but it seems sufficient.

Did you try reversing the leads one side of the transformer? Maybe it's some kind of phantom phase coupling problem. Of course I don't have the slightest idea what I'm talking about.

The reason you're experiencing this problem is due to the A/D chip used in the DPM. It's likely the 7106 DMP chip, which uses an internal voltage reference. The chip design is such that the low side of this reference must be maintained at a voltage that is above the analog common pin. <p>There are a couple solutions to the problem. The first, and most obvious, is to provide a separate supply to power the DPM. This can be either a separate transformer-operated supply, of an ISOLATED DC-DC converter that can be run from the supply that you want to monitor. You may be able to find a suitable DC-DC converter from the listings at http://www.jameco.com/webapp/wcs/stores ... ryId=11650<p>A second solution is to modify the circuitry on your DPM to add an external reference to the circuit and not use the internal reference.<p>Another solution is to build a circuit to provide a negative supply for the A/D chip. This can be run from the supply that you're monitoring also.<p>More detailed info on all the above suggestions can be seen in the datasheet and application notes on the 7106 chip from the Intersil.com web site (http://www.intersil.com/products/device ... pn=ICL7106).<p>Cheers!!!!!

Using the isolation transformer should work. But let me ask a couple questions first.<p>Did you hook the primary of the isolation transformer to the secondary of the power transformer and NOT the DC section of the power supply? Then did you feed the secondary of the isolation transformer to a SEPARATE bridge rectifier and not simply feed the regulator with AC? Then did you feed the DC from the bridge rectifier to the capacitor and regulator? I know these questions sound stupid but I'm just trying to prompt you to see that you didn't fall into a GOTCHYA that sometimes happens. Also, you will want caps on both the input to the regulator and to the OUTPUT as well.<p>Try these tests. Disconnect the meter so that the power supply for the meter is OPEN, i.e. no connections made to the ISOLATION power supply. Using a multimeter (and not one of the panel meters) measure the output of the DC regulated supply for the meter using both AC and DC measurements. Idealy this would be made with a scope but can be made with a digital multimeter. Make sure you don't have ripple or funky voltages. My guess is that you may not be drawing enough current with just the meter as a load to make the regulator function properly. I don't recall what the minimum load needs to be but something like 5mA should do the trick. The meter itself will only draw in the microamps and might be at fault. Since you are using a 9 volt regulator a 1.8k Ohm resistor loading the "meter" power supply will draw 5mA. But this is just an assumption and you might need slightly more current.

De-energize the power supply and discharge all of the capacitors. Use an Ohm meter to confirm that there is no DC path back through the transformer and no windings are jumpered or shorted. <p>Barring all of that you could build a completely separate power supply, with separate transformer, dedicated to the meter. Or you could even kludge some sort of wall wart to the inside of the power supply for supplying power to the meter. Not real clean but cheap if you use surplus wall warts.<p>Now for other theories and/or possibilities. I am throwing these out for critique and welcome any suggestions, corrections, or information.<p>How about using a pair OP AMPS for a virtual ground and supply to supply the meter? Using a dual op amp chip wire one op amp as a virtual ground and the other as a comparator so its output is high. Power the OP AMP using the DC section of the main power supply. Connect the meter to the outputs of the two op amps, observing proper polarity. Use a regulator, if needed, using the outputs of the OP AMP. I would think if you are using a power supply that has constant, regulated base, but feeds an adjustable section, you could power the OP AMP from the constant section. The gain of the "positive" OP AMP could be adjusted using a pair of resistors to supply the proper output.<p>Since the meter requires so little current, connect a non-polarized capacitor to each leg of the main power supply transformer secondary. Connect each of the other legs of the two capacitors to the inputs of a bridge rectifier. Filter the output of the bridge rectifier accordginly and connect it to the meter. Thus the ground to the meter is isolated by blocking capacitors. (Attempting ASCII schematic below)<p> --------------||----------
Transformer........................Bridge Rectifier
--------------||---------<p>
Just some thoughts. Comments anyone?<p>[ January 25, 2005: Message edited by: Jolly Roger ]</p>

From what you said about when it works, and when it doesn't; there IS a connection between the commons. Maybe the cases of LM317 and LM7809.