DM42- clock losing time

Post here to share useful tips and tricks, to ask questions about using your DM42 or to report software-related problems
ctrclckws
Posts: 173
Joined: Sun Feb 17, 2019 4:30 pm

Re: DM42- clock losing time

Post by ctrclckws »

Received my DM42 in August of 2018 so it's pretty new S/N 3108.
No real issues with time keeping.
So maybe the design has been tweaked in more recent batches?
DM10, DM10L: 00031 / DM11, DM11L: 00112 / DM12, DM12L: 02074
DM15, DM15L: 11069 / DM16. DM16L: 02001 / DM41, DM41L: 00859
DM41X: 00036ß / 00181 DM42: 3108 / 6084 WP43: 0032
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: DM42- clock losing time

Post by Walter »

Jebem wrote:
Fri May 10, 2019 8:04 pm
I don't remember to see any adverts from the manufacturer concerning the precision of the clock but i might be wrong.
Nor do I. Though it's not a matter of specification or advertising, it's just a matter of customer expectation. If and when you include something like an RTC in a calculator nowadays, a customer will expect it showing the real time. And this may be off for seconds but not for minutes IMHO.

ctrclckws wrote:
Fri May 10, 2019 9:57 pm
Received my DM42 in August of 2018 so it's pretty new S/N 3108.
No real issues with time keeping.
So maybe the design has been tweaked in more recent batches?
Maybe. But one correct device may be a lucky chance hit as well. Nobody knows but SM.
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
TwoWeims

Re: DM42- clock losing time

Post by TwoWeims »

I tune my HF rig to WWV and let my DM42 listen to it over night... :lol:

You just have to inject some humor into this very serious discussion.
cappy
Posts: 93
Joined: Sun Jul 30, 2017 7:30 pm

Re: DM42- clock losing time

Post by cappy »

TwoWeims wrote:
Sat May 11, 2019 1:44 am
I tune my HF rig to WWV and let my DM42 listen to it over night... :lol:

You just have to inject some humor into this very serious discussion.
I love the Weimars, I’ve been around them most of my life.
DM42 SN: 0612
cappy
Posts: 93
Joined: Sun Jul 30, 2017 7:30 pm

Re: DM42- clock losing time

Post by cappy »

ctrclckws wrote:
Fri May 10, 2019 9:57 pm
Received my DM42 in August of 2018 so it's pretty new S/N 3108.
No real issues with time keeping.
So maybe the design has been tweaked in more recent batches?
In my case, it lost time while the dm42 wasn’t in use.
It may be something to do when the calc is turned off as opposed to being used very often.
DM42 SN: 0612
ctrclckws
Posts: 173
Joined: Sun Feb 17, 2019 4:30 pm

Re: DM42- clock losing time

Post by ctrclckws »

Mine is also mostly in my work bag, waiting for its time in the sun.
DM10, DM10L: 00031 / DM11, DM11L: 00112 / DM12, DM12L: 02074
DM15, DM15L: 11069 / DM16. DM16L: 02001 / DM41, DM41L: 00859
DM41X: 00036ß / 00181 DM42: 3108 / 6084 WP43: 0032
User avatar
Jebem
Posts: 21
Joined: Sun Jul 23, 2017 12:21 am
Location: Portugal

Re: DM42- clock losing time

Post by Jebem »

Walter wrote:
Fri May 10, 2019 11:08 pm
Jebem wrote:
Fri May 10, 2019 8:04 pm
I don't remember to see any adverts from the manufacturer concerning the precision of the clock but i might be wrong.
Nor do I. Though it's not a matter of specification or advertising, it's just a matter of customer expectation. If and when you include something like an RTC in a calculator nowadays, a customer will expect it showing the real time. And this may be off for seconds but not for minutes IMHO.
I see you point.

However the installed crystal features 20ppm, resulting in around 60 seconds per month drift.
So the worst case scenario for the reported clock issues should be around 6 minutes after 6 months.
Crystal tolerances and temperature variations explain the other better looking reported values.

To have a proper precise RTC significant effort and cost would be required during the design stage, starting with very stable and expensive components selection (crystal and load components like capacitors and resistors) .
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: DM42- clock losing time

Post by Walter »

OK, 20ppm corresponds to 10.5 minutes per year (that's far worse than any cheap quartz watch sold nowadays). But now we must talk about the 20ppm - how do you get them?
  1. Are these crystals produced according to this spec in an industrial process? I hope so. Then, to warrant almost all crystals meeting this spec, standard deviation (desvio padral) of the production process (following a normal distribution) shall be less than 20ppm/6 = 3.3ppm. The crystals would come out distributed with 95% or more of them between -10 and +10ppm.
  2. If, OTOH, these crystals are picked out of a population having a significantly greated SD (due to a less precisely controlled production process), you'd need a sorting machine featuring a precision of ~1.3ppm and expect the crystal accuracy being uniformly distributed between -20 and +20ppm.
Finding out more about it would require some time accuracy statistics: how many DM42 show which deviation from real time?
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
User avatar
Jebem
Posts: 21
Joined: Sun Jul 23, 2017 12:21 am
Location: Portugal

Re: DM42- clock losing time

Post by Jebem »

Walter wrote:
Sat May 11, 2019 11:03 pm
OK, 20ppm corresponds to 10.5 minutes per year (that's far worse than any cheap quartz watch sold nowadays). But now we must talk about the 20ppm - how do you get them?
These are regular electronics components found in shops. Nothing fancing like the ones used in professional and lab equipments where the oscillator is temperature controlled and tolerances are tight.

Some time ago i created a thread in the other forum where i tried to detail the hardware components used in the DM42. I have used a picture posted by someone in the net and the crystal chip is clearly visible. By checking the crystal specifications from the reference code printed on the chip Q1 we kniw it is a 20ppm component:

https://www.hpmuseum.org/forum/thread-1 ... light=Dm42


20ppm on a crystal is just the regular low precision part. 10ppm or less will be more expensive.
But as i said, the precision is not only dependent on this component. The load circuit is critical to maintain the oscillator frequency as close as possible to the specification. For instance, bad quality capacitors will cause frequency deviation with temperature.
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: DM42- clock losing time

Post by Walter »

Bom dia Jose, y muito obrigado for pointing to your comprehensive investigation. I admit I've read it a year ago but didn't remember. BTW, please note you called the quartz a "high precision crystal" in your post of 2018-02-11 and didn't relativize this statement any later AFAICS.
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
Post Reply