Affordable 4-digit panel meters?

I'd appreciate knowing your sources on the 3 1/2 digit panel meters that you have first hand experience. Basically overall quality of construction, did they ship any duds, stuff like that.

There is a "PM328" DPM that is available for under $20:

Perhaps this is the manufacturer:

I would say that the $20 meters you found were probably not "standard pricing" but were probably no-names.

Very very few real-world applications can use the extra digits usefully. There are some customers that really need it, but most customers demanding more digits are on an ego-trip :-).

The demand for standalone meters seems to be getting smaller and smaller as everyone moves to microcontrollers and commodity A/D that actually matches the resolution/accuracy needed. If you need isolation then you have to think a bit. In a lot of one-off applications it doesn't pay to think, you just buy the standalone meter!

Tim.

Want to replace the analog meters in my bench PS with digits. It goes beyond

20, so I'll need 4-digit meters, yes? I'd like more than 1 decimal place.

Maybe I've got it wrong... How do these meters work? Do they show the max decimal places possible (

Typically the surplus 3.5 or 4.5 digit DPM unit itself is either 0-2V full scale, or 0-200mV full scale.

Some units can be jumpered for either, and some other ranges (including current ranges) are available. If you buy from a distributor you will have much better choice of ranges compared to buying a surplus unit.

You are responsible for adding the dividers/shunts if you want, and you are responsible for putting the decimal point where you want.

If you want 0.1V resolution, then a 3.5 digit meter gets 0.1V through

199.9V. Or at 0.01V resolution, a 3.5 digit meter gets 0.01V through 19.99V. As commonly incorporated into low-end power supplies, you might find a slide switch that changes the divider and the decimal point.

I would seriously question the need (other than "ego trip") of going to a 4 digit meter. Especially if (for example) you aren't buying 0.1% or better resistors for your divider or shunt already.

You also may have to look at where the power to the DPM module is coming from if you do not have a separate isolated power source just for the meter and you need to float the meter reference above ground. One often-seen option is a battery.

Tim.

"Tim Shoppa" wrote in news: snipped-for-privacy@o5g2000hsb.googlegroups.com:

If you are powering from the same regulator that powers a sensitive circuit that is outputting millvolts for measurement, take care, especially if you're using an LED meter. The meter can change the load, the load regulation might allow the other circuit's output to change, in turn changing the meter reading...

The easiest way to avoid this really vicious cycle is to use separate regulation for the meter and the measured, or even to isolate them with separate transformer windings, if battery power is not important.

There may be other, better advise to come on this subject. I hope so, cos I'll be watching it avidly if there is.

innews: snipped-for-privacy@o5g2000hsb.googlegroups.com:

All the DPM meters I've ever dealt with have separate connections for power ground and signal ground. (On Modutec brand meters, they're "V-" and "IN LO", respectively.) Tying them together at the meter is (as you pointed out) a Bad Thing - so you run them back to your common ground. Right?

Tim.

"Tim Shoppa" wrote in news: snipped-for-privacy@o5g2000hsb.googlegroups.com:

That's how I did it, problem still happened. DPM is made by 'Summit', and specifies a direct (or via 10K) connection between Gnd and in- if the meter and measured are otherwise isolated from ech other, but it doesn't specify what to do if they aren't. I put it down to an LM2904 LDO 5V regulator not having the load regulation I was hoping for. I'd had the circuit on a pinboard previously and powered by a simple smoothed supply and LM317 to regulate for 5V, and I guess the large caps in the external supply and the LM317 together prevented the problem that arose with later supplies based on a small 12V SMPU and the LM2905.

If the meter and measured share a power supply, and the measured is making a local ground after being supplied by small DC-DC power converter, I guess it complicates things. I ruled out the power converter as being the cause though, I think. What I found was small shifts in voltages all over the system in sync with the LED meter's updates, if the value changed.

It was actually interesting, in a perverse way, sometimes steady, sometimes oscillating between two values, or two different values, or three, or a random series, it was like an illustration of chaos theory.

innews: snipped-for-privacy@o5g2000hsb.googlegroups.com:

You'd have to look up the meter's common-mode rejection, but I'd rather find a signal common for "IN LO", and not even tie it to power ground, just signal ground (which might or might not be in common with power ground - it depends on the circuit). If it's a decent meter, they "should be" totally isolated (up to its CMR, of course.)

If it's a cheap crapola meter, then all bets are off. :-)

Cheers! Rich

I plan on installing completely separate supplies (xfmr, regs, etc.) for each meter.

I've not decided on LED or LCD.

Each meter will do double duty as V and I meter (via toggle switch).

My problem is understanding my need for accuracy.

The outputs are 0-25/0-25/0-16. So my presumptions are: - a 3-1/2 digit meter that is wired for 20v max will give 2 decimal places of accuracy; wired for 200v max it will give 1 decimal place of accuracy. - a 4-plus digit meter will give an additional digit of accuracy under these conditions.

So I must decide if 0.1v accuracy is OK for 2 of the outputs.

Comments welcome.

Thanks,

--
Al, the usual

Usual Suspect wrote in news: snipped-for-privacy@news.sf.sbcglobal.net:

Two. You'd be feeding it through a potential divider to get the 2V or 200 mV the meter wants. The accuracy would be the same if your dividers were good. You'd move the decimal point when selecting your divider for range.

A 200 mV input meter might accept 20V while indicating over-range, so a

200V line could safely be connected to a /10 divider. Just don't let it switch to the meter input directly in error. >:)

The dividers are important, that's how you control the accuracy, the meter then shows the most significant digits, and your dividers can be hand- picked resistors giving better accuracy than the meter's own.

That's true, even with the significant digits thing solved, 25.00 wants 4 full digits. Don't forget that there's no point in the extra digit unless the meter's native accuracy supports it. The accuracy and maximum count of

99.99 is 5 fold higher than 19.99, so a tripling of cost isn't a bad spend.

[...]

It's also possibly conversion noise from the meter getting back into the circuit being measured, not through the meter power lines, but through the meter sensing line. Typically the conversion cycle is exactly tied to the LED update rate.

I have seen some perverse cases where crystal oscillators in my circuit intermittently lock to EMI coming from switching power supplies etc in prefab modules I was using. In that case it was radiated, not conducted, EMI.

Tim.

Way, way overkill, man. It sounds to me like you're going to be building 3 power supplies to support your upgraded existing power supply.

Watch out for voltage drop across toggle switch and shunt, and now you'll have to do current shunts too.

Yeah, you're the one who has to decide that.

What's the resolution/accuracy you got from the old analog meters?

0.1V is 0.4% of your 25V full-scale, and that's pretty good for a big old analog meter with a mirrored scale. More typical accuracies for a small front-of-power-supply analog meter is 2 or 3%.

And don't confuse resolution with accuracy. You'll have to be using better-than-1-percent resistors (probably just go to 0.1% resistors) in your dividers and shunts etc.

This is beginning to sound more like an exercise in specmanship (well, not even that, more like an exercise in how many digits can we display) than any actual utility.

If you're getting the impression that I actually prefer analog meters for many quantities, you'd be correct! Do you really need to know the current to a fraction of a percent, or just see if it's drawing any at all or pegging the needle? Some of us know exactly what sort of overload it is by not looking at the analog meter, but by the noise it makes when it hits the peg!

Tim.

"Tim Shoppa" wrote in news: snipped-for-privacy@y80g2000hsf.googlegroups.com:

Interesting, but doubtful, it would have to go back through a /10 divider pair of resistors, then into the output of an op-amp, and it never happened on the pinboard layout with different supplies. I'll look into it with that in mind next time I try stuff, but I still think poor load regulation is the usual suspect. Sorry, that phrase seems to be haunting me, lately. :) If the problem turns out to be particularly vicious and mysterious, I might even start calling it Keyser Soze...

"Tim Shoppa" wrote in news: snipped-for-privacy@l77g2000hsb.googlegroups.com:

I like 1%, it's cheap, metal film is stable. I hand-pick mine. I buy 10 of

910R, ten of 8K2, I measure each 910R with a Fluke 79 (or whatever is the best meter in reach). I note the values in a list, then repeat for the 8K2. I multiply the fist real '910R' value by 9, and if there is a perfect match in the 8K2 list, I have my pair. If not, I move to the second 910R value...

Doing this is a cheap way to get at least a couple of resistors matched for a /10 network. If the resistors were tighter in tolerance it might be harder if not impossible to find pairs this way. 1% similarly allows enough likely excursions to give you a /100 pair of 100R and 10K. Both pairs can have accuracies of around 0.1% or better. They're unlikely to be the weak link in a chain of accuracy.

You also need to remember that a DC meter will read the average, and not the true RMS (AC+DC). When you have a a power supply that has significant ripple and noise, the reading may not be accurate for your purposes, and the additional digit of resolution will not be meaningful. 25.0 volts out of 199.9 will still be better than 1/2% of reading for 1 count.

Paul

Another problem, in addition to the metering resolution and accuracy, is the resolution of the power supply's adjustment controls. Can you actually set the output to 10mv resolution on the 0-25V range? If not, then there's no reason to expect that a higher resolution meter will make it happen. Connect a 4- or 4-1/2 digit DMM to the output and see if you can accurately and repeatedly set the output to within 10mv of your desired voltage. Unless your power supply has good coarse and fine voltage setting controls, I bet you can't.

--
Dave M
MasonDG44 at comcast dot net  (Just substitute the appropriate characters in the 
address)

Life is like a roll of toilet paper; the closer to the end, the faster it goes.

you

All very good points I hadn't considered. Many thanks to all who contributed.

Truly grateful for this forum.

Thanks,

--
Al, the usual

I have some PM128 3 1/2 digit panel meters for sale @ $12.00 + shipping.

Listing on my web site at:

Ed, W1AAZ