I just finished my boxes. The refs are isolated in RF boxes with some capacitors on Vin then those boxes are soldered shut(two spots to make sure shielding is continous. Those boxes are put into some Hammond boxes with a wire soldered to the RF box for earth connection and closed up. My reading offsets dropped by 50%. Was -40uV from nominal, down to -20uV.
Hello,
do you have a cirquit diagram of those references?
do you supply them by a switchmode power supply.
Is there any switch mode power supply (or LED lamp) near the cirquit?
What happens with battery supply?
In principle I see two possible issues:
- either instability due too much capacitive loading if the output has no measure for isolating capacitive loads.
- or instability due to EMI/switchmode supplies.
Sealed? No ventilation? I've read in this very thread (somewhere, can'f find it now--sorry!) that no ventilation is bad because the LTZ1000 will lose regulation at some point...
but not at room temperature. Most of the heat is transported by the component leads and by the copper on the PCB. So the chance of over isolation is very small.
I suggest to do a measurement of T.C. then you can see where the regulation begins to be weak (non-linear) and when it is completely failing (with 40-50 ppm/K).
with best regards
Andreas
I just finished my boxes. The refs are isolated in RF boxes with some capacitors on Vin then those boxes are soldered shut(two spots to make sure shielding is continous. Those boxes are put into some Hammond boxes with a wire soldered to the RF box for earth connection and closed up. My reading offsets dropped by 50%. Was -40uV from nominal, down to -20uV.
Hello,
do you have a cirquit diagram of those references?
do you supply them by a switchmode power supply.
Is there any switch mode power supply (or LED lamp) near the cirquit?
What happens with battery supply?
In principle I see two possible issues:
- either instability due too much capacitive loading if the output has no measure for isolating capacitive loads.
- or instability due to EMI/switchmode supplies.
Sealed? No ventilation? I've read in this very thread (somewhere, can'f find it now--sorry!) that no ventilation is bad because the LTZ1000 will lose regulation at some point...
but not at room temperature. Most of the heat is transported by the component leads and by the copper on the PCB. So the chance of over isolation is very small.
I suggest to do a measurement of T.C. then you can see where the regulation begins to be weak (non-linear) and when it is completely failing (with 40-50 ppm/K).
with best regards
Andreas
He sent me full schematics and parts lists for both. I've attached the buffered output section which is what I use to watch it. I've been powering them with a DP832 that they sit on top of and ALL the lighting is LED lighting.
So if I read this right, your reference is outputting 10V +/- whatever the multiplied deviation on the LTZ1000 is from 7.14V? Not a bad way to go, as long as you know the voltage. No pots to drift (I like things a little over 10V, rather than a little under).
These output about 3-4 uV under raw zener output. 213 is ~7.090466V. I was considering adding a 10V buffered output but I'm not that worried about it now.
He sent me full schematics and parts lists for both. I've attached the buffered output section which is what I use to watch it. I've been powering them with a DP832 that they sit on top of and ALL the lighting is LED lighting.
And where are the capacitors that you added?
Only on power supply or also on output?
with best regards
Andreas
He sent me full schematics and parts lists for both. I've attached the buffered output section which is what I use to watch it. I've been powering them with a DP832 that they sit on top of and ALL the lighting is LED lighting.
And where are the capacitors that you added?
Only on power supply or also on output?
with best regards
Andreas
I only added the capacitors to the input of the circuit across the feed throughs. So inside the hammond box but outside the RF box. I should also mention I mistakenly purchased mini banana plugs for the output so my testing has only been done with poorly fitting leads with the boxes. New standard banana jacks and new end plates are on their way and I'll start logging consistently again.
I only added the capacitors to the input of the circuit across the feed throughs.
So no capacitor at the output and no ferrite or something like that -> it is most likely conducted EMI.
You can check this if the output changes by moving wires or holding the isolated wires by hand.
If the output changes then it is 100% EMI.
Unfortunately the output buffer will get instable with large capacitive loads according to my simulation.
So you could try 1 to 5 nF at the output or a R/C low pass with 47R - 100R and 100nF at the output.
In this case you will have to switch the DMM to > 10 GOhm mode.
with best regards
Andreas
I only added the capacitors to the input of the circuit across the feed throughs.
So no capacitor at the output and no ferrite or something like that -> it is most likely conducted EMI.
You can check this if the output changes by moving wires or holding the isolated wires by hand.
If the output changes then it is 100% EMI.
Unfortunately the output buffer will get instable with large capacitive loads according to my simulation.
So you could try 1 to 5 nF at the output or a R/C low pass with 47R - 100R and 100nF at the output.
In this case you will have to switch the DMM to > 10 GOhm mode.
with best regards
Andreas
Even at 2s(minimum integration time for 7.5D) I only see the lsd change with twisted 1 meter long leads. The output leads are twisted to the glass isolated pass through, then twisted again to the binding posts. The meter leads are also twisted from the box to the meter. The only change I saw from no box to RF and aluminum box is the 20uV offset change. I assume this is from better thermal isolation though, no breezes and more gradual temperature changes.
I'll swap the end plates to move to standard bananas and monitor for another 36 hours to see if any other changes took place. The ltz1000A board was ±0.4ppm and the ltz1000 was ±2pm but the 1000 was bothered more by my kids during the day so those readings weren't great. I can't actually grab the soldered wires any more but I don't remember having issues with that before. Once I get new data I'll see if it seems to need these alterations.
Test your reference with the lid on and off, see if the change is immediate with the absence of the lid. You could have a switching mode supply somewhere near (or not even that near) which is causing RFI/EMI and the chopper amp is rectifying affecting your output. I just about went crazy with this exact same issue. The steel Tekno boxes will work very well to shield both RFI/EMI.
Hello,
modifications on my new LTZ1000 (non A) references:
LTZ#7 calculating around 40 ppb for each 1 Meg Resistor for R9 gave 3 in parallel so 330K to compensate for 120 (138) ppb.
LTZ#8 calculating also around 40 ppb gave a 910 K resistor for R9 to compensate for 45 ppb.
Measurement Result:
LTZ#7 gave +62 ppb/K in average. So in realty we had +67 ppb/1 Meg resistor.
LTZ#8 gave -5 ppb/K in average. So this has around +36 ppb / 1 Meg resistor.
Measures:
LTZ#7 changed R9 = 510K instead of 330K and soldered the inner housing which was only loose up to now to ease component changes.
LTZ#8 I am unshure wether I should change R9 = 820 K instead of 910K.
(here I am already within my usual +/- 20 ppb/K target).
with best regards
Andreas
Hello,
further modifications:
LTZ#7: now with R9 = 510K and inner housing soldered as planned. -> -10 ppm/K over 18-38 deg C
LTZ#8: As I had to solder the inner housing I also changed R9 = 820K -> +10 ppm/K in average.
with best regards
Andreas
Not quite the same level of detail as Andreas, but I've been logging one of my LTZ1000A's over the past year. Its surprisingly stable, according to my Simpson 8.5" needle 260, its not budged at all
Who needs an 8588A anyway.
On a serious note, I found this the best solution to simply monitor the output while it cooks for long term aging. Need something to show the output is at least present, and its very energy efficient
When a company I worked for acquired Sola Electric, a colleague of mine found Joseph Sola's Simpson 260. Joseph Sola patented improvements to the Ferroresonent Transformer, a.k.a. Constant Voltage Transformer in the late 1930's. His Simpson 260 had a very low serial number, single digit or teens, IIRC.
Relevance- Interesting how mankind used to be able to measure voltage without a precision reference!
Hello,
LTZ#7 in final adjust state with R9 = 475K
Between 18 and 38 deg C the average linear T.C. is -2 ppb/K
Box T.C. is 0.92 ppm/20 deg C so 46 ppb/K in the limited temperature range.
Above 38 deg C the T.C. increases so the overall box T.C. is 66 ppb/K.
For my needs the reference has to work to about 33 deg C so there is also some reserve.
I will have to re-test the power consumption as the temperature setpoint has changed.
with best regards
Andreas
Is it 33°C environmental, or internal temperature?
Your cabinet upstairs is getting really hot in summer, isn't it?
basement is much better!
Frank
Hello Frank,
unfortunately the range 18 deg (winter) to 33 deg (last summer)
environmental temperature is that what I have as "normal" in my "lab".
And that with up to 6-7 deg C change over one day.
So I am really envious on your lab in basement.
with best regards
Andreas
That's not too bad, it's 10C to 38C here. Up to 18C change in a day. Looking at getting a mini split put in so I can keep it MUCH cooler and stable.
Andreas,
so what is the temperature rise from environmental to internal temperature?
If I remember right, mine is about 7°C , or so. I did not isolate the LTZ1000 against the PCB, but have a styrofoam layer completely around the PCB
Couldn't you arrange a small lab place at the level of your houses entrance?
Frank
so what is the temperature rise from environmental to internal temperature?
Hello,
Of course the difference depends on environmental temperature
(as the heater power decreases with environment temperature increase).
Attached you find some temperatures of the measurement above.
Environment (of LTZ) temperature (measured on top of the aluminium housing of LTZ#7)
PCB temperature measured inside LTZ#7
Heater temperature of the heat spreader (aluminium sheet) below the aluminium housing of the reference within the car cooling box.
ADC25 temperature (the temperature of the measurement system outside the car cooling box).
with best regards
Andreas
Arrived 2 sets of precision metal foil resistors for LTZ1000
The manufacturer promises very good characteristics of these resistors. So will check it.
@ bsw_m
could You write approximate prices of these devices ?
Price is not so high. Write pm.
Price is not so high. Write pm.
Is this a special secret or can you share price, source and specs to everyone?
-branadic-
Price is not so high. Write pm.
Is this a special secret or can you share price, source and specs to everyone?
-branadic-
I think it will make sense after testing
Are these hermetic? They don't look so.
