Control.com - Nerds in Control

I need a paperless recorder that can record Instruments analog signals (4 to 20mA) and digital signals (from field to DCS and vice versa) with a very short sampling time (1ms or less) so that we have first out alarm in equipment trips scenarios. We prefer this recorder to have clamp pins so that we can connect around the wires which are already connected to DCS and do not affect the process by removing these wires.
The objective is that we need to put a trap for the suspected trip signals for equipments such as pumps or compressors that sometimes trips with many signals come to DCS at same time and we can not identify what happened first as we do not have the fist out alarm facility on our DCS. It is preferred to be portable compact size so that we can take it to any system. Can any one help??

You need a Digital Storage Oscilloscope with a decent memory size. Trigger off of your alarm or whatever event occurred and hook up to all the signals you suspect might be the problem. We use HP scopes with 2 analog channels and 16 digital channels and what they call "Megazoom" (Extra capture memory so you can record many seconds and then "zoom in" on the event that caused the problem. You didn't mention the time period you want to capture, so I'm thinking you want a finite number of seconds before and after your error events. If you want minutes or hours than the scope might not be a good device as this gets into the realm of a data logger device.

I would wire some short leads off of your device (for voltage probing). You might need to add a resistor to the 4-20ma loop and use a differential probe, unless you have a current probe that is a DC clamp on type of probe (I've used these, they work well but aren't cheap). There might be other tricks.

There probably exists a breakout Pod that works over USB to a laptop that will do what you need. A scope is overkill in some ways, but in other ways you can see if you have high frequency noise riding on your signal whereas some of the cheaper alternatives might cap out at 100Khz or something like that (Which is still pretty nice for industrial uses, don't get me wrong). My point is that if you do any electrical work at all an investment in a good multi-use oscilloscope from name brands (HP, Tektronics, Lecroy) is not a bad thing at all. The digital inputs aren't needed much, but they are really handy when you need to use them as a poor mans timing analyzer to look at a handful of signals. I think if I didn't have an 8-16 digital input accessory for a scope I would want to get at least a 4 channel scope.

I've been assuming you can get at all of these signals within a reasonable distance. Matters get complicated if all of these devices don't come back to a central area. This problem will get more complex if that is the case.

KEJR

DAQ,

I haven't shopped 'high speed' data acquisition in 2 decades but I'm inclined to agree about a digital storage scope to get a 1KHz sampling rate. I'm familiar with paperless recorders and I am unaware of anything faster than 50 Hz (for 8ch) sampling rates, and that's with very iffy triggering.

The Rigol DS1102E 100MHz 2 ch storage digital scope has recently dropped in price here in the US from ~$750 to $400 recently. It is very favorably reviewed on YouTube. It's 2 channels with a 3rd trigger channel. I just bought one, it's very nice little box.

The normal memory depth can be 8K, the extended long memory depth 512K.

At 1KHz, that's 8 seconds or 512 seconds duration. If you're looking to distinguish events in the 1mS range, wouldn't 8 seconds be more than enough memory depth?

I suspect an independent trigger channel might be important, in order to 'trap' your signal. There is some basic math like ch1-ch2, as well.

The vertical voltage is a typical 2mV to 10V/div on a 1/2/5 sequence. There is no real-time clock correlation to acquired data, so if you need data from multiple scopes, there's no means of time synching. I believe Rigol (and others) have 4 channel models.

Probes:
Fluke's 77x line is handheld clamp-around-probes for 4-20mA circuits. The high end 773 appears to have an analog output, but nowhere does the spec mention frequency response. My gut feeling is that for a handheld the frequency response is tied to the LCD display update rate, but who knows, 10x/sec? I somehow doubt the output has a 1KHz response.

There's the AEMC model K100 clamp-around DC current probe, but it is clear that it measures AC and DC components, which Fluke doesn't mention, but I'll bet it does nonetheless. If the 4-20 signals are HART, the 1200 FSK will appear riding on top of the DC signal. At a 1Khz sampling rate, 1200 baud will alias. I'm not sure it matters, unless it affects triggering. It claims a 2KHz freq response.

All DCS's I've heard about use a 250 ohm shunt resistor at the analog input. I'd be inclined to get out the marshaling panel wiring diagrams and hunt down the terminal strip location and clip a standard scope probe across the resistor to get the 1-5Vdc signal.

The Rigol is a scope so the std probe ground is chassis ground referenced. If the presence of a ground on a DCS input alters the signal in the control room, then float the scope with a cheater on the AC power line, I won't tell anyone. Plus there's likely to be a 110Vac duplex around for power to run the scope.

A lot of 'dataloggers' use a single amp and a single A/D and multiplex all their numerous channels.

The resulting skew between channels can be significant, particularly if time relationship is critical.

Shop carefully. Caveat Emptor

David

The recording time may be for 24 hours or less or more and sampling rate within 1m second. I do not know is the major instruments vendors like Honeywell, ABB, Yokogawa,....etc. can have such a portable compact recorder??

It's called a Digital Storage Oacilloscope. Get one with a deep memory and decent trigger facilities and you can capture lots of pre and post trigger information. One that I've used extensively for this type of problem is:

http://www.amazon.com/Rigol-DS1052E-50MHz-DSO/dp/B003MYND5A/ref=pd_ sbs_indust_1

The clamp bit will take some ingenuity as current loops are fussy. The scopes are really cheap now and a fantastic bargain for the capability if you know how to use one.

Regards
cww

24 hours at a 1KHz sampling rate is 85 mega words/channel,

Check out the white paper here:
http://www.dataq.com/applicat/articles/hsda.htm
where the author talks about streaming to disk at high rates. Maybe Dataq has something.

Be sure to inquire about software used to look through that much data. I wouldn't want to scroll through it.