| Electronics > Projects, Designs, and Technical Stuff |
| DIY DC Current Probe? |
| << < (13/13) |
| 0xdeadbeef:
Crap. Just ordered another two CMS3025 ... just in case. |
| David Hess:
The old Tektronix way is to insert a hall effect sensor into the split ring used for a clamp-on AC current probe. The hall effect sensor does not measure the low frequency or DC current. Instead, it measures the flux balance, positive or negative, in the core and a DC current is driven into the sense winding (or a dedicated winding) to reduce the DC and low frequency flux through the core to zero. By using the imperfect hall effect sensor to balance the flux to zero, its poor linearity is not a factor. Holding the low frequency flux through the core at zero also prevents saturation of the core at high DC currents. Combining the signal driven into the balance winding with the signal from the sense winding produces a frequency response from DC to 10s of MHz; 40 or 50 MHz is typical. |
| capt bullshot:
That's exactly the way the typical LEM (and other brands) current sensor works. Except they use an iron core and a secondary winding of 1000 or 2000 turns, resulting in a bandwidth of 500kHz (ballpark). The Sensitec ones use an GMR sensor that is most sensitive to the U-shaped conductor on the bottom side of the ceramic substrate. Within the sensor itself, the working principle is the same, compensating the external field by driving a smaller current through an on-chip U-shaped conductor. They do not have a magnetic core, so they're more sensitive to stray magnetic fields. So their sensing accuracy mostly relies on mechanical tolerances of the U-shaped conductor and the chip placement. They also have these little bias (permanent) magnets required for the MR sensor to work, if these get demagnetized or mechanically damaged (one can tear them off the substrate with mild forces), the sensor doesn't work anymore. Afaik there's little to none magnetic coupling (transformer) effect within the Sensitecs, so the amplifier loop has to provide for all the measurement bandwith - that's most probably the reason why they could do the CMS3000 (higher BW than CMS2000). LEM and others have a simpler low BW loop amplifier (as the Tek clamps do) and must gain their BW from the transformer. That's their limit in BW. In terms of DC accuracy (offset and temperature stability), the LEM ones perform better than the Sensitec ones. I know some people who called the Sensitec CMS2000 series thermometers, anyway, there was another Sensitec series (CDS afair, 5V operated) that had a significant long term drift. There's another technology, Flux Gate by VAC, also compensating to zero flux. These are quite good in DC accuracy, but rather noisy from my experience. Their Flux Gate operating frequency (the main source of the "noise") is within the measurement bandwidth, so one cannot easily filter them. There's a big market for this kind of sensors: Inverters and VFD / Servo drives. They don't require that large BW, the usual 500kHz is more than sufficient here, so I never saw the point for the CMS3000 series. |
| Jeroen3:
--- Quote from: 0xdeadbeef on November 05, 2018, 07:25:44 pm ---Crap. Just ordered another two CMS3025 ... just in case. --- End quote --- 25A is within the range of the CFS1000, 5A is not. --- Quote from: capt bullshot on November 06, 2018, 06:28:23 am ---In terms of DC accuracy (offset and temperature stability), the LEM ones perform better than the Sensitec ones. I know some people who called the Sensitec CMS2000 series thermometers, anyway, there was another Sensitec series (CDS afair, 5V operated) that had a significant long term drift. --- End quote --- I have to disagree with you. LEM CASR sensor are noisy crap, maybe other types are better but my overall impression of LEM is poor. If you have any other current carrying wire, screwdriver or metal object near 5 cm of these things, these drift away and become unusable. God forbid you using a magnetized screwdriver on the connector near them. We have had to encapsulate them in ferrite tape to get stable readings. |
| mk_:
--- Quote from: Jeroen3 on November 05, 2018, 05:57:58 pm ---Sensitec discontued the CMS current sensors. Back to square one. Or if I can aqcuire some to build a few units, since the prototype has been extremely useful. --- End quote --- I had a discussion with one of their guys on electronica about this as we are using those parts in quantities. Today I received an official answer that they build every order coming in until december, expected shippingdate at the moment is week 25/2019... hth |
| Navigation |
| Message Index |
| Previous page |