Hi.

If anyone's considering buying one of those expensive Salter Electronic
Kitchen Scales, you'll probably want to be aware of a design problem
that messes up the accuracy. You can work around the problem, and still
get accurate readings, but it is a rather daft fault. This is what I've
put on Salter's feedback form, and I'm waiting for their reply:

-------------------- Begin quote --------------------

I bought one of your 1004 "Electronic kitchen scales with stainless
steel platform" about a year ago and I was very happy with it. However,
it developed a fault, and I had it replaced under guarantee with a new
one. Unfortunately, I am far from happy with the behaviour of the new
scale and I have found what I believe is a serious design fault that has
been introduced into it.

I noticed that when I was measuring powders or liquids, that I was
ending up with quantities that were visibly too large for the displayed
weight. When I then transferred the powder or liquid to some old
balance scales that were cumbersome but accurate, I found that the
quantities were as much as 20% too high.

I then investigated the problem, and found that the scale misreads if
the powder or liquid is added too slowly. I presume that this is because
a software change has been introduced into the scale's electronics to
try to compensate for drifting errors in the weight sensors.

Didn't it occur to the designers that the natural way to use a kitchen
scale is to put a bowl on the scale, zero it, pour in most of the
quantity of powder that is required, then slowly add the rest to bring
it up to the target weight. If you do this however, the scale
under-reads. The only way to get around this behaviour, and to get a
reasonably accurate reading, is to press a spoon into the powder for a
moment, when adding small quantities, to make a large change to the
weight which disables this software trick.

I have to say that this has made the scale much harder to use and is
very irritating. I'd like to return the scale for one that behaves the
way that the old scale did. Before I do however, I'd like to know if any
are available, or do all of your kitchen scales now have this misleading
drift compensation thing?

-------------------- End quote --------------------

--
Farry

On Sun, 25 Jan 2004 10:16:35 GMT, Farry
wrote:

| Hi.
|
| If anyone's considering buying one of those expensive Salter Electronic
| Kitchen Scales, you'll probably want to be aware of a design problem
| that messes up the accuracy. You can work around the problem, and still
| get accurate readings, but it is a rather daft fault. This is what I've
| put on Salter's feedback form, and I'm waiting for their reply:

I use the Salter electronic Aquatronic model number 3007, all plastic, and
find it everything I had hoped. I have checked the accuracy a few times
and found it fine.

Dave F

Dave Fawthrop wrote:

I use the Salter electronic Aquatronic model number 3007, all plastic, and
find it everything I had hoped. I have checked the accuracy a few times
and found it fine.

I would also have said that, but of the scale I bought a year ago,
rather than the "compensated" scale. When did you buy yours?

--
Farry

On Sun, 25 Jan 2004 10:38:41 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| I use the Salter electronic Aquatronic model number 3007, all plastic, and
| find it everything I had hoped. I have checked the accuracy a few times
| and found it fine.
|
| I would also have said that, but of the scale I bought a year ago,
| rather than the "compensated" scale. When did you buy yours?

Sounds as if yours has a simple mechanical fault. Perhaps crud has built up
inside, or the platform has got twisted.

I have had mine about ?2? years.

I have just tried adding sugar slowly, no problem.

Dave F

Dave Fawthrop wrote:

On Sun, 25 Jan 2004 10:38:41 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| I use the Salter electronic Aquatronic model number 3007, all plastic, and
| find it everything I had hoped. I have checked the accuracy a few times
| and found it fine.
|
| I would also have said that, but of the scale I bought a year ago,
| rather than the "compensated" scale. When did you buy yours?

Sounds as if yours has a simple mechanical fault. Perhaps crud has built up
inside, or the platform has got twisted.

I have had mine about ?2? years.

I have just tried adding sugar slowly, no problem.

OK, the model 3007 is not displayed on their website, so I guess it's an
old model and predates the compensation trick.

As I said, it's the brand new scale that's got the problem. It's almost
certainly designed into the software because I can think of no way that
a fault could introduce behaviour like that, but I can think of why a
software designer would want to add that compensation, if he didn't
properly think through the consequences.

--
Farry

On Sun, 25 Jan 2004 11:12:35 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| On Sun, 25 Jan 2004 10:38:41 GMT, Farry
| wrote:
|
| | Dave Fawthrop wrote:
| |
| | I use the Salter electronic Aquatronic model number 3007, all plastic, and
| | find it everything I had hoped. I have checked the accuracy a few times
| | and found it fine.
| |
| | I would also have said that, but of the scale I bought a year ago,
| | rather than the "compensated" scale. When did you buy yours?
|
| Sounds as if yours has a simple mechanical fault. Perhaps crud has built up
| inside, or the platform has got twisted.
|
| I have had mine about ?2? years.
|
| I have just tried adding sugar slowly, no problem.
|
| OK, the model 3007 is not displayed on their website, so I guess it's an
| old model and predates the compensation trick.
|
| As I said, it's the brand new scale that's got the problem. It's almost
| certainly designed into the software because I can think of no way that
| a fault could introduce behaviour like that, but I can think of why a
| software designer would want to add that compensation, if he didn't
| properly think through the consequences.

Just reread your original post.

As both an ex-engineer and a software writer, it still looks like a
mechanical problem, and not a software problem. My guess is that the
platform is dragging, catching or juddering on something.
Mechanical problems sometimes build up slowly, or suddenly appear.
Software problems are either there, or not there, all the time.

Dave F

"Farry" wrote in message
...
Dave Fawthrop wrote:

On Sun, 25 Jan 2004 10:38:41 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| I use the Salter electronic Aquatronic model number 3007, all plastic,
and
| find it everything I had hoped. I have checked the accuracy a few
times
| and found it fine.
|
| I would also have said that, but of the scale I bought a year ago,
| rather than the "compensated" scale. When did you buy yours?

Sounds as if yours has a simple mechanical fault. Perhaps crud has built
up
inside, or the platform has got twisted.

I have had mine about ?2? years.

I have just tried adding sugar slowly, no problem.

OK, the model 3007 is not displayed on their website, so I guess it's an
old model and predates the compensation trick.

As I said, it's the brand new scale that's got the problem. It's almost
certainly designed into the software because I can think of no way that
a fault could introduce behaviour like that, but I can think of why a
software designer would want to add that compensation, if he didn't
properly think through the consequences.

--
Farry

The kitchen scales use strain guages as the weight transducer. Although
these are not linear in output (which is compensated), they do not exhibit
the fault you describe. Its more likely to be a fault in the weighing
platform suspension mechanism sticking (imagine a shaft going through a hole
and rubbing on the side). Probably a batch manufacturing problem. Check the
batch number of your model and take it back for exchange, making sure the
new one is from a different batch/manufacture date.

Dave

Dave Fawthrop wrote:

Just reread your original post.

As both an ex-engineer and a software writer, it still looks like a
mechanical problem, and not a software problem. My guess is that the
platform is dragging, catching or juddering on something.
Mechanical problems sometimes build up slowly, or suddenly appear.
Software problems are either there, or not there, all the time.

OK, I've just done this.

1. Place clean mirror on worktop to make a perfectly flat surface.

2. Make sure both sides of scale are clean, and place on mirror.

3. Zero scale.

4. Put cereal bowl on scale - 274g.

5. Add tablespoon (20g) of muesli - 294g.

6. Tip out muesli and put bowl back - 274g.

7. Add same muesli back very slowly over 1 minute - 276g.
i.e. it's somehow lost 18g.

8. Tip out muesli and put bowl back - 256g.
Still 18g down.

9. Remove bowl - and scale shows "----" (below zero).

This is quite repeatable. The scale is accurate and consistent, and does
not drift even over several minutes, PROVIDED that the weight changes
are swift.

QED?

--
Farry

On Sun, 25 Jan 2004 13:08:45 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| Just reread your original post.
|
| As both an ex-engineer and a software writer, it still looks like a
| mechanical problem, and not a software problem. My guess is that the
| platform is dragging, catching or juddering on something.
| Mechanical problems sometimes build up slowly, or suddenly appear.
| Software problems are either there, or not there, all the time.
|
| OK, I've just done this.
|
| 1. Place clean mirror on worktop to make a perfectly flat surface.
|
| 2. Make sure both sides of scale are clean, and place on mirror.
|
| 3. Zero scale.
|
| 4. Put cereal bowl on scale - 274g.
|
| 5. Add tablespoon (20g) of muesli - 294g.
|
| 6. Tip out muesli and put bowl back - 274g.
|
| 7. Add same muesli back very slowly over 1 minute - 276g.
| i.e. it's somehow lost 18g.
|
| 8. Tip out muesli and put bowl back - 256g.
| Still 18g down.
|
| 9. Remove bowl - and scale shows "----" (below zero).
|
| This is quite repeatable. The scale is accurate and consistent, and does
| not drift even over several minutes, PROVIDED that the weight changes
| are swift.
|
| QED?

Judder!
A well known mechanical engineering problem. Often experianced in car
brakes when the car vibrates when braking.
Friction varies with position and time, not like they teach you in school.
I have seen it in lots of things but never in scales.

Take it back and demonstrate the problem, Even a shop assistant should
understand that it does not work properly. Get a different model, it may
be a batch problem.

Dave F

Farry wrote:
6. Tip out muesli and put bowl back - 274g.

7. Add same muesli back very slowly over 1 minute - 276g.
i.e. it's somehow lost 18g.

8. Tip out muesli and put bowl back - 256g.
Still 18g down.

9. Remove bowl - and scale shows "----" (below zero).
QED?

Interesting. Some recipes call for a half dozen ingreditents and I add them
one a t a time. It can easily take a few minutes as you put one away and
get out the next. That would be a problem.

--
Ed

http://pages.cthome.net/edhome

Dave Fawthrop wrote:

Judder!
A well known mechanical engineering problem. Often experianced in car
brakes when the car vibrates when braking.
Friction varies with position and time, not like they teach you in school.
I have seen it in lots of things but never in scales.

Take it back and demonstrate the problem, Even a shop assistant should
understand that it does not work properly. Get a different model, it may
be a batch problem.

I can't take it back to the shop, because I bought the original a year
ago, and now I've just got a replacement under guarantee by post from
Salter. So I'll wait for the response from Salter to my query, first.

I appreciate that it's hard to believe that this behaviour could be
designed in. OK, next experiment.

I repeat the process of adding muesli to the bowl three times, without
zeroing the scale, and weigh the bowl each time. It starts at 274g, then
it's 254g, then 236g, then 216g. i.e. 18-20g down each time leaving it
58g low. And since I completely remove the bowl from the scale to tip
out the muesli each time, that should have sorted out any judder.

--
Farry

"Edwin Pawlowski" wrote:

Interesting. Some recipes call for a half dozen ingreditents and I add them
one a t a time. It can easily take a few minutes as you put one away and
get out the next. That would be a problem.

Actually, that can be a problem, but not for the reason that I've
outlined. If you leave the scale without changing the weight on it, the
display will switch off after a couple of minutes to save power, and you
can't switch it on again without zeroing the scale. You have to remember
to dab it with your finger every minute to keep it alive, and the
displayed weight will remain consistent. Just don't pour in an
ingredient too slowly.

--
Farry

"Dave Gibson" wrote:

The kitchen scales use strain guages as the weight transducer. Although
these are not linear in output (which is compensated), they do not exhibit
the fault you describe. Its more likely to be a fault in the weighing
platform suspension mechanism sticking (imagine a shaft going through a hole
and rubbing on the side). Probably a batch manufacturing problem. Check the
batch number of your model and take it back for exchange, making sure the
new one is from a different batch/manufacture date.

Unfortunately, there's no batch number, the model number is the same,
and the appearance of the old and new scales are identical, as far as I
can remember. I believe that I've eliminated the possibility of the
mechanism sticking mechanically, so that leaves either a bizarre
electronic fault, or more likely in my opinion, an ill-thought attempt
to compensate for drift in the strain gauge.

--
Farry

On Sun, 25 Jan 2004 16:27:49 GMT, Farry
wrote:


| I can't take it back to the shop, because I bought the original a year
| ago, and now I've just got a replacement under guarantee by post from
| Salter. So I'll wait for the response from Salter to my query, first.

Salter offer a 10 year guarantee
http://www.salterhousewares.com/page...x.asp?code=14#
and click on 1004.

Why not get it swapped for a 3007 which they still sell on
http://www.salterhousewares.com/page...e=14&offset=2#
Not as fancy but it works OK.

Better than stripping what you have and risking buggering it up.

--
Dave Fawthrop
Sick and tired of Junk Snail Mail? Register your family
surname and address with www.mpsonline.org.uk

Dave Fawthrop wrote in
:

On Sun, 25 Jan 2004 13:08:45 GMT, Farry
wrote:

| Dave Fawthrop wrote:
|
| Just reread your original post.
|
| As both an ex-engineer and a software writer, it still looks like a
| mechanical problem, and not a software problem. My guess is that
| the platform is dragging, catching or juddering on something.
| Mechanical problems sometimes build up slowly, or suddenly appear.
| Software problems are either there, or not there, all the time.
|
| OK, I've just done this.
|
| 1. Place clean mirror on worktop to make a perfectly flat surface.
|
| 2. Make sure both sides of scale are clean, and place on mirror.
|
| 3. Zero scale.
|
| 4. Put cereal bowl on scale - 274g.
|
| 5. Add tablespoon (20g) of muesli - 294g.
|
| 6. Tip out muesli and put bowl back - 274g.
|
| 7. Add same muesli back very slowly over 1 minute - 276g.
| i.e. it's somehow lost 18g.
|
| 8. Tip out muesli and put bowl back - 256g.
| Still 18g down.
|
| 9. Remove bowl - and scale shows "----" (below zero).
|
| This is quite repeatable. The scale is accurate and consistent, and
| does not drift even over several minutes, PROVIDED that the weight
| changes are swift.
|
| QED?

Judder!
A well known mechanical engineering problem. Often experianced in car
brakes when the car vibrates when braking.
Friction varies with position and time, not like they teach you in
school. I have seen it in lots of things but never in scales.

Take it back and demonstrate the problem, Even a shop assistant should
understand that it does not work properly. Get a different model,
it may be a batch problem.

Dave F


Not convinced Dave. At least it must be coupled with a software fault
if the bowl showed different weights at different times.

--
Adrian