Reducing digital scale drift

[Rocketvapor]

@SPJ
Is the 20.03 displayed known error in your check weight and the 20.05 display drift from the 20.03?

I use something a little bigger than 20 grains.
My charges are about 30 grains, so a 100 grain or 200 grain Fake Zero shows enough zeros.
For example (the 200 grain bullet is good) Put 100 grains, scale shows 100.00, add powder it shows 130.00.
Dump pan replace on scale shows 100.00. Scale factor @ 100 is good, zero at 100 is good.
But you guys practice and see if it works and holds zero better.

My A&D scale should be here Friday and the cheap ebay scale Saturday.
Oh Boy, more experimenting

 

[Rocketvapor]

Be careful NOT to blame check weight errors on CALIBRATION.
With a good zero, a Full Scale calibration should only be required when conditions change.
Scale moved or temperature changed. For a 50 or 100 gram scale with milligram resolution, error will likely be percentage based. If the cal factor is stable to 20ppm/0.002% per degree F, you will not see it at small loads. Changing position can cause larger errors.

With a good zero, cal @ 100 grams (or 50grams if required).
Let the scale sit at zero for a few minutes to relax the load cell.
Now weigh your cal weight, and suppose it shows 10 counts low,
49.990 (-0.02%) or 99.990 (-0.01%). At smaller loads, like a few grams of powder, that calibration error won't be seen (maybe a count? ). Unless you are OCD like me, don't fret a few counts at heavy load.
Zero, then Fake Zero stability, will matter much more than calibration error.

The 300 gram scale posted can be calibrated @ 200 grams, or 100, 200, 300 grams.
The 3 point cal reduces linearity errors between zero and 100 grams.
Design a check method for YOUR scale with good check weights.
The 1, 11, 111 gram check I did was a fast and easy method of checking Cal factor, linearity, and zero.

 

[JFrank]

@SPJ
Is the 20.03 displayed known error in your check weight and the 20.05 display drift from the 20.03?

I use something a little bigger than 20 grains.
My charges are about 30 grains, so a 100 grain or 200 grain Fake Zero shows enough zeros.
For example (the 200 grain bullet is good) Put 100 grains, scale shows 100.00, add powder it shows 130.00.
Dump pan replace on scale shows 100.00. Scale factor @ 100 is good, zero at 100 is good.
But you guys practice and see if it works and holds zero better.

My A&D scale should be here Friday and the cheap ebay scale Saturday.
Oh Boy, more experimenting

I have two of those 20. Gr weights that actually weigh 20.2 along with the 100 gr for calibration I just wanted to understand the fake zero thing.
Half hour later the damn scale drifted more then ever before.

 

[Rocketvapor]

That needs to be investigated a little more.

 

[RickKr]

Be careful NOT to blame check weight errors on CALIBRATION.
With a good zero, a Full Scale calibration should only be required when conditions change.
Scale moved or temperature changed. For a 50 or 100 gram scale with milligram resolution, error will likely be percentage based. If the cal factor is stable to 20ppm/0.002% per degree F, you will not see it at small loads. Changing position can cause larger errors.

With a good zero, cal @ 100 grams (or 50grams if required).
Let the scale sit at zero for a few minutes to relax the load cell.
Now weigh your cal weight, and suppose it shows 10 counts low,
49.990 (-0.02%) or 99.990 (-0.01%). At smaller loads, like a few grams of powder, that calibration error won't be seen (maybe a count? ). Unless you are OCD like me, don't fret a few counts at heavy load.
Zero, then Fake Zero stability, will matter much more than calibration error.

The 300 gram scale posted can be calibrated @ 200 grams, or 100, 200, 300 grams.
The 3 point cal reduces linearity errors between zero and 100 grams.
Design a check method for YOUR scale with good check weights.
The 1, 11, 111 gram check I did was a fast and easy method of checking Cal factor, linearity, and zero.

I'm a little confused by use of different units, ala references to "count" and "ppm". I noticed your earlier reference to a 300 / ±0.001 gram scale having 300,000 counts. Is this just another way of expressing the overall range and resolution? And how does ppm come into this?

And I haven't even tried to understand the temperature factor, but because my shop is so stable I'm not sure it is much of a factor for me.

I had not considered doing anything to check linearity, mostly because I wasn't aware of a way to do it. Looking my scale's manual, I see it does show linearity calibration masses of 50 and 100 grams. There is nothing there, however, about how to do that. I will attempt to find out from Ohaus.

That does bring to mind that I do not know the real world accuracy of the 100 gram standard that came with the unit, and certainly do not know the same for the 10 and 50 gram standards that came with the Hornady scale I have. However, given what you say above about a few counts off at the calibration weight won't really be seen at powder weights (5 to 70 grains in my case), maybe investing in higher accurate weight standards is not really warranted. I have considered taking what I have to a pharmacy or lab to get them weighed on something appropriate, but...

And lastly, I may be a bit confused about when to use the "zero" function. I have been operating under the assumption it should not be used immediately following calibration with the 100 gram weight. My assumption being that doing so changes things inappropriately relative to the calibration. I note with context that it might be appropriate to wait a few minutes with the scale resting at zero following a calibration to let the load cell relax.

Rick

 

[JFrank]

That needs to be investigated a little more.

It’s a piece of crap like most of budget electric scales.
My tuned beam scales kick its ass and never drift

 

[RickKr]

I have two of those 20. Gr weights that actually weigh 20.2 along with the 100 gr for calibration I just wanted to understand the fake zero thing.
Half hour later the damn scale drifted more then ever before.

That needs to be investigated a little more.

Is there any air movement in the room, heating/cooling/fans units running? I have a ceiling heater and a split-type heat pump and a circulating fan running in my shop, which I have to turn off all of in order to have the air be still during weighing. If I don't, I see large excursions of the scale readings. My scale does have a small hinged cover to help reduce this, but it is not convenient to use during powder weight because it doesn't close over the film canister I like using.

Rick

 

[Kracken McGraw]

It’s a piece of crap like most of budget electric scales.
My tuned beam scales kick its ass and never drift

SPJ just out of curiosity, pull that clear lid off and put it aside and see if that makes any difference on drift.

 

[JFrank]

SPJ just out of curiosity, pull that clear lid off and put it aside and see if that makes any difference on drift.

Brother I’ve been snowed in for a month, and retired sooo WTH I’ll give it a shot and report back.
I’m all fairness that’s the most drift I’ve ever seen out of this little scale, normally a few hundredth is about max.

 

[Kracken McGraw]

Brother I’ve been snowed in for a month, and retired sooo WTH I’ll give it a shot and report back.
I’m all fairness that’s the most drift I’ve ever seen out of this little scale, normally a few hundredth is about max.

I have read somewhere about removing the lid might help and do it on mine, but it may just be all BS.

 

[Rocketvapor]

The big plastic lid might be acting like a wind flag.
Or, could be static. Try it without the lid and see.
Maybe a big cardboard box with an access cut out made from an Amazon box
Your scale may be the problem.
Maybe a wrist strap, anti-static mat, power conditioner might help.
Seems necessary for some of the more expensive scales.
If your can't get stable readings, something is going on.
Even the $20 digitals perform better than that. I wouldn't trust a digital that you can't PROVE to be stable with powder charges. Beam scale might be necessary in your situation.

Since MOST digital scales background NATIVE unit of measurement is GRAMS/MILLIGRAMS, accuracy will be a little better than having the scale calculate and round off readings for GRAINS.
That's one of the advertising points of the Creedmoor Sports new scale.
Native unit is GRAINS. But, it's not that difficult to convert loads in grains to gram/milligram.
A milligram is a little less than 0.02 grains. +/- 6 milligrams is about 0.1 grain (some simple rounding here).
+/- 3 counts on a milligram indicating scale (about 0.05 grains) is about the BEST you can expect with a tuned beam scale. That's probably good enough for almost everyone.

The 300 gram scale I've shown doesn't seem to be available now. I don't know if the look alikes offered now are any good. Having such a large range it is NOT good to trickle small powder charges.

The units I've used in this discussion may seem foreign to some. Sorry.
Counts, percent of reading, percent of full scale, ppm.
For a milligram indicating scale, a count is one milligram. In grain mode a "count" may be 0.02 grains.
A 20 gram, milligram indicating scale has 20,000 counts from a zero to a full scale of 20.000 grams,
a 100 gram, milligram indicating scale has 100,000 counts from a zero to a full scale of 100.000 grams.
Make sense? The 300,000 counts, from zero to 300.000 grams seems unbelievable for a cheap scale.
Got lucky I guess.
At high loads, over 250 grams to near 300 grams, I normally see 5 or more (maybe 10) counts of repeatability.
At 300.000 grams, +/- 5 counts is +/- 0.005 grams, +/- 0.0017%, or +/- 17ppm.
At a powder charge of 2.000 grams, I can get to +/- 2 counts fairly easily using a Fake Zero.
That is +/- 0.002 grams (+/- 0.031 grains), +/- 0.1%, or 1000ppm. (I think I got those correct).

I have a 50 gram scale that indicates to 0.05 grain, 0.002 grams, but to 0.005 carats.
0.005 carats IS 0.001 gram. Go figure

This fake zero thing isn't going to work for everyone. Give it a try. If your scale drifts, figure out why.
(or toss it out the window )

 

[Rocketvapor]

I've collected several cheap scales (don't believe in the buy once cry once method ).
Some OK, some not.


Got the stove back.
Here are some junk Chinese 10 gram weights from a big box of culls.
Across the back of the scale is one that is 5mg light, 4 pretty close, and 4 heavy, by 20 to 30 mg.
The others are 3 of my standards, a Troemner and a Rice lake and a very old Christian Becker.
Sequence was S1, T1, T2, S2, T3, T4, S3.
Normalized the 4 test weight results to an average correction of -0.5mg.
Fake Zero stayed good within a count.



Didn't bother with the S.D. but bet it's less than 2 counts.
When my A&D gets here I'll repeat this and adjust the slightly high ones down.
Found 3 more worth checking. 7 test weights and 3 standards would give me 100g in 10g steps to compare with my 100g standard.

The ten gram test weights seem to range from -1mg to +1mg. That's about as close as I can figure with this scale.

The 101 gram reading has been bouncing between 101.000 and 101.003 today.

 

[Rocketvapor]

Enough with this scale. The EJ-54D2 was just delivered.
Might do a review after a couple of days.
Not sure I like the plastic draft shield.
It allowed a 20 gram calibration on the low range.

 

[Rocketvapor]

Still have some more learning to do with this scale and the various menu options before I start a review.
Turned Zero Tracking OFF, Slow sensitivity ON, Stability sensing to 1/2 count.
Ran it overnight and cold (68F@ 5AM) the zero drifted about 3 mg, displayed 0.0032. Performed a Zero, and cal factor @ 10 grams was surprising. See pic below. Heater is on, house warming up.
A quick check of Standard S1 and S3 (see previous post comparing 10 gram weights on 300 gram scale) shows S3 is lighter by 0.4mg.

 

[brokeasajoke]

I've collected several cheap scales (don't believe in the buy once cry once method ).
Some OK, some not.
View attachment 1311454

Got the stove back.
Here are some junk Chinese 10 gram weights from a big box of culls.
Across the back of the scale is one that is 5mg light, 4 pretty close, and 4 heavy, by 20 to 30 mg.
The others are 3 of my standards, a Troemner and a Rice lake and a very old Christian Becker.
Sequence was S1, T1, T2, S2, T3, T4, S3.
Normalized the 4 test weight results to an average correction of -0.5mg.
Fake Zero stayed good within a count.
View attachment 1311456

View attachment 1311483
Didn't bother with the S.D. but bet it's less than 2 counts.
When my A&D gets here I'll repeat this and adjust the slightly high ones down.
Found 3 more worth checking. 7 test weights and 3 standards would give me 100g in 10g steps to compare with my 100g standard.

The ten gram test weights seem to range from -1mg to +1mg. That's about as close as I can figure with this scale.
View attachment 1311637
The 101 gram reading has been bouncing between 101.000 and 101.003 today.

I have one of those little Frankford Arsenal scales I use a lot but it's problem is drift while I have the pan off dumping a charge. Maybe a fake zero would help it. I have a FX120i and a Parker 10-10 but that little dude is just so handy when I'm only putting together 2 or 3 rounds.

 

[JMayo]

Don't breath on it !

 

[Rsadams]

Interesting read and thanks for posting.... I learned here in Texas that we have wild temp swings , so wild that it of course interferes with the house temperature.... If the temp changes the digital scale needs to be recalibrated.... I realize alot of shot shells and them being plastic can really get the scale infested with static , which leads to crazy weights and drift....

 

[Dave M.]

Good Morning
@ 6AM this morning, temperature in the kitchen was 67F.
Scale had "Drifted" to 5.003. Added 100 grams and got 104.992 grams (with 105.005grams loaded on the scale).
Zero seems to have drifted down another count and Full Scale Calibration down about 13 counts or about 0.01% .
More than I expected. Heater is on, will check again once house warms up.
EDIT: 10AM
House now at a toasty 72F, 71F under scale base.
5.007 (+2 counts) / 105.001 (-4 counts, -6 counts including zero shift)
1PM
Full scale hasn't come back. Still 6 counts low with 105.005 grams on the pan, 5.005 grams 2 counts high at 5.007. Did a three point calibration @ 100, 200, 300. Starting off with just the 100 gram weight. Getting 99.999. Added 1 gram, 100.999. Another cold (cold for us) night. Conditions not right for reloading, might as well let the scale sit.

Conditions are ALWAYS right for reloading.

 

[Dave M.]

Yes, I've noticed that about several of the supposed check weights, off by 0.01-0.04 grns. The 50 grn weight is actually from a Hornardy scale kit, not the Lyman. My intent in using them is for reference and consistency, not accuracy. That the fake zero is not exactly 100.00 grn is as you say, for monitoring the process. I could tell I overshot it slightly when making it, as it consistently weighed between 99.98-100.00. I figure if I keep it on the scale when measuring powder charges and disregard the 100 digit, I'll be more consistent than when relying on the auto zero function of the tared canister. And as you say, my confidence is already higher

Rick

Is all of this info still relevant while using a magnetic force restitution scale like the A&D FX120i? Just curious
Dave

 

[Rocketvapor]

Dave,
The Fake Zero is a method to monitor zero drift that MIGHT be hidden with an AUTO ZERO.
Unless you have the option to turn off auto zero it could be a function of any voltage digitizing device.
If you find it necessary to continuously re-zero, watching a Fake Zero value might help.
Try it and see if your Fake Zero drifts. I used 10 grams so there would be plenty of zeros for weighing charges.
A Force restoration scale measures the current supplied to the coils with a fixed number of turns (ampere turns) by measuring voltage with an AtoD converter.
There is a scale that uses a capacitor to control an oscillator and measure frequency.


The 120i has this in the specs:
"Compact SHS (Super Hybrid Sensor) with 1 Second Stabilization"
The specs list Repeatability (1 S.D.) as 0.001gram, Does that include zero drift caused by external sources, electrical noise, temperature?
Linearity as 0.002 grams. Does not affect readings over a limited range verified with check weights,
and Sensitivity Drift of 2ppm/Deg C, I interpret that as cal factor drift. (that's pretty darn good).
Sounds pretty good. In the proper environment this would be a great performer.

Maybe someone could try it with a dispenser by adding a fake zero check weight to the pan and to the charge value.

Just a reminder.
MOST scales weigh things more precisely in their native unit. That's usually grams not grains.