|
Post by Lake Shore Ryan on Oct 26, 2023 15:35:47 GMT -5
Hi Dmitry,
Honestly, it's hard to say with the wire. If you have a lot of RF noise sources in your lab (like large motors) beyond what is present in a traditional cryogenic system (things like turbopumps aren't that noisy), then you will be fine with ribbon wire. The added thermal connection this wire provides to heat sink bobbins will be worth it.
The grounding issue is a deep topic. By suggesting to connect shields to the cryostat only, we're trying to avoid ground loops, which result in 50/60Hz signals flowing along the shields and potentially coupling into the sensor wires. Suggestions for both the 336 and 372 although they may use slightly different approaches, are all trying to avoid this issue. The instrument will be connected to an electrical ground, which may be at a different potential to the cryostat, which will likely be connected to a different earth ground. For your shield to be effective, it must be connected to one of these grounds. The cryostat is what we suggest. The main takeaway though should be to never connect both ends of the shield on the cable. If you connect one end to the cryostat, and the other end to the instrument ground, you could end up with a worse situation than no shield at all. Current will flow between the two potentials, likely with a 50/60 Hz component from the power lines which can couple into the signal wires.
Hope this helps.
|
|
|
Post by Lake Shore Ryan on Oct 18, 2023 8:11:28 GMT -5
Hi Dmitry, Good catch. I can see how this could be interpreted this way. As with most things, there is nuance to recommendations. Here is my explanation for why we made these product choices. In general, we find that noise inside the cryogenic environment is minimized due to it being inside a grounded metallic enclosure. We found that poor thermal anchoring is the greatest source of measurement error in user systems, so we chose to use ribbon cable to maximize thermal contact when wrapping the wire around a thermal anchor. The ribbon cable we use on temperature sensors with maximum temperatures of 420 K or less uses a bonding agent to create the ribbon cable. This bonding agent would melt before reaching the maximum rated temperature of the diode (500 K), so we instead use twisted pairs for this CU-HT package. This wire is rated for use to 500K. Also, diodes aren’t expected to go as cold as Cernox sensors, so the best possible thermal anchoring isn’t as critical as if it was a Cernox sensor attached to something reading below 1 K. We felt this was a reasonable trade-off. Outside of the cryostat, we absolutely recommend using shielded twisted pairs for all sensor types regardless of the wire inside the cryostat. Hopefully this helps explain our seemingly inconsistent product offering. If your application would benefit from twisted pair wiring on the Cernox CU-HT package, there are several options: - You can cut the wire right back to within a few centimeters of the sensor and attach your own twisted pair wiring. This won’t change the sensor calibration curve as it is wired in a 4-lead configuration.
- We can customize the sensor for you for an additional cost. This usually becomes economical when dealing with many sensors as the cost of customization is amortized over many sensors. You can request more information about this by contacting your local sales representative.
Thanks for your post Dmitry.
|
|
|
Post by Lake Shore Ryan on Jul 5, 2023 10:44:56 GMT -5
Hi labmaster, do you have two separate heaters providing heat to the same location where the sensor is installed?
|
|
|
Post by Lake Shore Ryan on Jun 5, 2023 12:59:32 GMT -5
For anyone who finds this in the future, we've now improved the website. It does a slightly better job letting people know about these option cards.
|
|
|
Post by Lake Shore Ryan on May 31, 2023 9:02:51 GMT -5
OK, no problem. This probably isn't a solution for now (unless you REALLY want all your measurements in the same Chart Recorder window), but the 3062 option card for the 336 expands the number of input channels up to 8. Sorry, I'm realizing now that we don't do a very good job advertising this card on our website. Here are some product links:
With this card installed in the 336, Chart Recorder would recognize and display all 8 channels in a single window. These 4 additional channels use the measurement circuitry of the 4th input of the 336 and scan sequentially through any active channels. So you'll end up with 3 full-speed channels, and up to 5 channels with a shared update rate. With all 4 of the inputs on the 3062 option card active, you'd end up with an update rate of 2 per second per input (5 channels sharing a 10 per second update rate).
|
|
|
Post by Lake Shore Ryan on May 30, 2023 10:55:52 GMT -5
Hi Leo,
If you don't mind me asking, are you using two 336s because you have two separate systems, or are you using two to get more heater power into a single system? Thanks.
|
|
|
Post by Lake Shore Ryan on Feb 20, 2023 13:59:09 GMT -5
Hi Adam,
Good question.
The Model 224 is passively cooled as it has a low power draw compared to our controllers. So rack ventilation alone should be sufficient to cool stacked instruments.
Higher-powered instruments like our Model 336 temperature controller are actively cooled with an internal fan that pull air through the instrument from one side to the other. So even these instruments should be suitable for stacking on top of each other in a rack.
Are you asking because you're planning on stacking this many Model 224 monitors? Or have you already done this and are experiencing problems?
|
|
|
Post by Lake Shore Ryan on Jan 13, 2023 12:14:37 GMT -5
Happy to wait and hear back about your experience with another probe. I've also given the support team a heads up about your case, so they'll be ready if you reach out to them. They'll be able to accept your products back for testing on our calibration stations to confirm whether they're performing within specifications.
|
|
|
Post by Lake Shore Ryan on Jan 9, 2023 14:30:11 GMT -5
adamp, you wouldn't happen to have another probe that you can swap with this one would you? This problem may require an RMA for that probe so we can properly compare it with our calibrated systems and see if we can figure this out.
|
|
|
Post by Lake Shore Ryan on Jan 6, 2023 17:46:58 GMT -5
Thanks for this extra information. Sorry for the delayed response, was on vacation. I'll talk to some people here about what could be causing this and get back to you. Thanks for your patience.
|
|
|
Post by Lake Shore Ryan on Dec 20, 2022 13:04:56 GMT -5
The teslameters accuracy specs are all defined within a ±5°C range, so small fluctuations like you saw shouldn't be a problem. I looked up your probe, and transverse probes like this (FP-2X-250-TSxxxxxxx) all use the same temperature compensation table: Temperature (°C) | Gain factor | 0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
| 1.0014
1.001
1.0007
1.0003
1
0.9997
0.9982
0.9962
0.9951
0.9937
0.9925
0.9916
0.9907
0.99
0.9893
0.989
0.9886
0.9886
0.9886
|
To check whether this is being applied properly, could you please try the test again with probe temperature compensation disabled? You should still record probe temperature and plot it against field strength, just ask the teslameter not to apply any temperature compensation. The results of this test will likely tell us whether there is an issue with the temperature compensation curve, or maybe some issue with the how the calibration was written or stored. Thanks.
|
|
|
Post by Lake Shore Ryan on Dec 19, 2022 14:08:01 GMT -5
Hi, thanks for giving such detailed feedback. Could you please let me know the serial number of the probe so we can check the calibration file for the probe. Maybe that will give us some clues.
As for your original post, all three of your operating assumptions at the end are correct. However, our temperature compensation algorithm uses a gain table (many temperature/gain pairs) with linear interpolation between the points in the table used to determine the gain for any given temperature value. This allows us to deal with the non-linear nature of a Hall sensor's temperature dependence.
This table is not unique for every sensor though, it is set for each sensor type based on population testing completed prior to the product's release. it may be that this probe's actual temperature coefficient is further away from the population average than normal, resulting in the plot you've shown. Based on what we see from the calibration file, we can make suggestions on how this might be corrected.
Is the teslameter also being exposed to these temperature fluctuations?
|
|
|
Post by Lake Shore Ryan on Sept 26, 2022 11:04:49 GMT -5
That's great to hear! I didn't realize you had a programmer, but this was going to be our next suggestion if initialization didn't work. Haha, I have to admit I didn't know about the John Harrison clock, and now I do. Sounds like excellent company for the 805 to be in .
|
|
|
Post by Lake Shore Ryan on Sept 26, 2022 10:34:10 GMT -5
Hi Tim,
Great to hear from you again! Thanks for posting this data. It's at least good to see the sensor remaining functional during these high pressure tests. Did the sensor readings happen to return to baseline after being pressurized? Or were they permanent shifts? Thanks.
|
|
|
Post by Lake Shore Ryan on Sept 23, 2022 14:00:24 GMT -5
Hi Ken,
I think you probably have the analog output set to show the raw signal, which is really only good for troubleshooting when measuring DC. You could try changing the analog output mode in your teslameter settings to see an output level that has been demodulated and is more representative of the field being measured.
|
|