EEVblog Electronics Community Forum
Electronics => Metrology => Topic started by: martinr33 on December 03, 2017, 09:11:18 pm
-
For some of us, we just can't have a nice controlled ambient temperature. In my environment, the temp can go from about 60 to about 85 (there are extremes, but not so common in the SF Bay Area).
The 3458a does move around in these extremes. +/- 5 - 10 uV, and autocal can take some of that out - depending.
My question is, if I switched out the fan for a more powerful, but thermistor controlled, unit - would this help?
Anybody tried this?
-
No, a more powerful fan would probably make the internal 3458A temperature more sensitive to the room temp.
Something more useful would be using a temp sensor / processor to control the fan speed and regulate the 3458A temperature that way. Lady Heather uses this approach to stabilize a GPSDO temperature. You set the desired control temp to a value between the room temp and the device self-heating temperature. The PID loop can stabilize the temperature to less than 0.01C with 24 hour RMS values in the low micro-degree range.
You now have to be concerned with the possible effects of the fan controller and fan electrical noise on the 3458A.
-
I am thinking, a fan that already has a built in controller. At temps below 37, the fan will run slower, so the internal heating drives the internal temperature up. At temps above 37, the fan will run faster, cooling the unit a little better at the expense of more fan noise.
The existing fan already has an asynchronous switch in it.
-
I would not change the flow rate at all, because that might disturb the instruments internal equilibrium.. that influence is very difficult to estimate, as the analogue circuits are very complex. The instrument has these circuits distributed on top (DC modes customization) and on bottom (AC and A/D), and all work together regarding thermal state.
Also, I really doubt that the internal temperature can be held constant by simply regulating the fan.. especially due to this mentioned distribution of circuits top / bottom.. There's a very delicate temperature gradient between DCV analogue circuits (where "TEMP?" is measured), and the A/D ASIC, which obviously both determine the DCV gain constant CAL? 72. So, it's not only necessary to keep the temperature as constant as possible, but also this gradient..
Another minor aspect is, to keep the internal temperature as low as possible, to preserve the components lifetime (especially these high speed comparators).
If you want to regulate the temperature by the fan, you would have to set the temperature towards these maximum about 70°C internal temperature (55°C ambient).
Don't you have a basement, where the temperature is more constant?
Frank
-
I'm with Frank on this one.
-
Correct - Don't mess with the 3458a - speaking from experience. It's only going to get worse if you do. Control your environment better; that's much more cost-effective. We've got customers in the SF bay area that have no problem holding a couple deg C in the working environment with minimal HVAC expense, which is where you have to be for low PPM work anyway.
The same thing holds for LTZ changing heater ratio resistors on a working, well aged 3458a. I wouldn't ever do that - all that does is screw up the Vref for years to come as it de-stresses after soldering, and a lot of "standard" 3458a's are working to well within -002 option stability spec anyway (as evidenced by cal data). Going to a cooler LTZ heater is going to have a minimal, if not detrimental effect on those units.
-
I removed the filter sponge of cooling fan..
-
Anybody tried this?
Yes, with a Solartron 7081. There is even an empty space for a fan in the rear panel.
I would not change the flow rate at all, because that might disturb the instruments internal equilibrium.. that influence is very difficult to estimate, as the analogue circuits are very complex.
That is why I circulated the air only through the logic section (which produces most of the heat anyway).
The analog and logic sections were connected with wind shield covered holes, which allowed slow heat exchange.
The result was almost as good as keeping the whole meter in a temperature chamber.