2 channel CTR0/CTR1 ouput synchronously

I am using PXI-6259 under Measurement studio VC++ and trying to generate 2 counter output synchronously.
********** DAQmx ***************
//set output channels
(DAQmxCreateCOPulseChanFreq(CTR0_Task,"Dev3/ctr0ev3/ctr1","",DAQmx_Val_Hz,DAQmx_Val_Low,0.0,CTR_Freq,0.5);
My program is fine on single channel counter output nomatter CTR0 or CTR1. But it is not working on 2 channels setting. Is there any limitation on 2 channels counter output generation synchronously?
Please advise me thank you very much!

I noticed that your number of samples has been set to 1000 (continuous mode), hence a possible reason for the timout errors. I would suggest changing that to 1 sample (on demand) and see if that gives you any progress.
-Ciao

Similar Messages

Trigger ai reading ctr0 ctr1

I have an E series board.
Problem description:
I want to set a pulse on ctr0 this pulse triggers a puls on ctr1 with a specified delay, and puls on ctr1 triggers reading of ai channel.
I wil implement this with DAQmx in Ansi C, but I have resource problems on the board.
Is there a solution to perform an ai measurement with a specific delay after pulse on ctr0.

This is the nidaq code I have this must be converted to NIDAQmx is this possible
// reset counter1
GPCTR_Control(iDevice, ND_COUNTER_1, ND_RESET);
// use counter 1 as puls generator to trigger counter 0
GPCTR_Set_Application(iDevice, ND_COUNTER_1, ND_SINGLE_PULSE_GNR);
// write delay counter1, note this delay is not used.
GPCTR_Change_Parameter(iDevice, ND_COUNTER_1, ND_COUNT_1,ulLOWcount);
// write pulse length counter1
GPCTR_Change_Parameter(iDevice, ND_COUNTER_1, ND_COUNT_2,ulHIGHcount);
// set gpctr0 as output channel of pulse of counter1
// To output a counter pulse, you must call Select_Signal.
Select_Signal(iDevice, ND_GPCTR0_OUTPUT, ND_GPCTR0_OUTPUT, ND_LOW_TO_HIGH);
// reset counter2
GPCTR_Control(iDevice, ND_COUNTER_0, ND_RESET);
// use counter as triggered puls generator1
GPCTR_Set_Application(iDevice, ND_COUNTER_0,ND_SINGLE_TRIG_PULSE_GNR);
// set channel other counter output to be used as trigger
GPCTR_Change_Parameter(iDevice, ND_COUNTER_0, ND_GATE,ND_OTHER_GPCTR_OUTPUT);
// write delay after external trigger
GPCTR_Change_Parameter(iDevice, ND_COUNTER_0, ND_COUNT_1,ulLOWcount2);
// write pulse length
GPCTR_Change_Parameter(iDevice, ND_COUNTER_0, ND_COUNT_2,ulHIGHcount2);
// set gpctr1 as output channel of pulse of counter2
// To output a counter pulse, you must call Select_Signal.
Select_Signal(iDevice, ND_GPCTR1_OUTPUT, ND_GPCTR1_OUTPUT, ND_LOW_TO_HIGH);
// arm counter2 (delay counter)
GPCTR_Control(iDevice, ND_COUNTER_0, ND_PROGRAM);
// program ADC to wait for trigger
iStatus = DAQ_Config(iDevice,1,0);
//Configure DAQ to look for Start Trigger from GPCTR1 Ouput routed via PFI_0 //ND_GPCTR1_OUTPUT
iStatus = Select_Signal(iDevice,ND_IN_START_TRIGGER,ND_GPCTR0_OUTPUT, ND_LOW_TO_HIGH); // changed output to gate
iStatus = DAQ_Start(iDevice,iChan,iGain,piBuffer,ulCount,ND_INTERNAL_1_MHZ,2);
// arm counter 1 (puls generator)
iStatus = GPCTR_Control(iDevice, ND_COUNTER_1, ND_PROGRAM);
while ((iDAQstopped != 1) && (iStatus == 0)) {
iStatus = DAQ_Check(iDevice, &iDAQstopped, &ulRetrieved);
iStatus = DAQ_VScale(iDevice, 1, iGain, dGainAdjust, dOffset,ulCount, piBuffer, pdVoltBuffer);
// CLEANUP - Don't check for errors on purpose.
iStatus = DAQ_Clear(iDevice);
iStatus = GPCTR_Control(iDevice, ND_COUNTER_1, ND_RESET);
iStatus = GPCTR_Control(iDevice, ND_COUNTER_0, ND_RESET);

DC motor control via ctr1/ctr0 pwm output stops outputting after one loop through.

Working on this project to control an H-bridge controlled
motor using pwm output from ctr0/ctr1 for ccw and cw on the motor. A sensor detects where it is at and each cycle goes from black to white. For some reason, the motor only turns on once and then nothing comes out of the ctr outputs after that. Anytime the program is stopped and restarted it works for one cycle. Or if just the program is run at the skipcolor sub-vi. this can be run multiple time to control it properly. During the actual run it doesn't work. Any implementation device would be great. I am using the PCI-MIO-16E controller.
Programs attached for reference.
Attachments:
SkipColor1.vi ‏111 KB
MotorControl2.vi ‏99 KB
PROJECT 2a2.vi ‏18 KB

Hey suspencefull,
You are probably going to have to think about re-designing your code. Your pulse train is not continually generating because you do not have it in a loop. When you use the DAQ Assistant to create a continuous pulse train, LabVIEW will prompt you if you want it to automatically add the loop. You must have deleted it or chosen not to add the loop.
There will, however, be problems though when you add the loop back in. The SkipColor sub VI will not finishing executing until the loop is finished and therefore you will never move out of the first frame of your sequence. I would suggest redesigning your code so that these thing can occur in parallel.
Please let me know if you have further questions.
Regards,
Chris Delvizis
National Instruments

What are the counter channels on PXIe-6361 pin-out diagram

Hello,
I defined three counters channels using DAQ assistant for the PXIe-6361 board (ctr0, ctr1, ctr2). I did not see explicit labels on the pin-out diagram for the PXIe-6361 card that indicate where these counters are. Can anyone point me out what are the pins / labels on the pin-out diagram for these counters?
Thanks much
Solved!
Go to Solution.

Hi Larrage,
The best way to find the default pins is using the DAQmx Help file. You can reach this in MAX by clicking on Help>>Help Topics>>NI-DAQmx>>NI-DAQmx Help. In the help file, you can search for "Connecting Counter Signals" and select the link for the series card you have (X Series in this case).
In your case, the default counter output terminals for ctr0, ctr1, and ctr2 would be PFI 12, PFI 13, and PFI 14, respectively. The counter inputs (for counting edges) would be PFI 8, PFI 3, and PFI 0 for the same counters. You can change the input/output terminals for counters in your VI or MAX, but these are the defaults for the card.
Zach P.
Product Support Engineer | LabVIEW R&D | National Instruments

Max number of Messages using Channel API?

We have a 2-port CAN card (series 2) runing at 500kB. When we use channel API to setup 9 messages (about 20 signal-channels) at 100Hz update rate in a write task - the data is constant, set to repeat output last sample (therefore butffer size is not the problem).
Task Init-start gives error of too many channels too high speed. when we set only 8 messages at 100Hz or 9 messages at 90Hz, it works fine.
If we calculate based on 500kB and 8-byte messages (give enough overhead), CAN bus shall be able to transact at least 40 full-length (8-byte) messages at 100Hz.
there are two Rx messages at 50Hz on the bus. (totally only these 10 or 11 messages are on the bus).
so what is the problem, simple because the CAN channel API is not good enough?

The NI-CAN driver I used was 2.3.1f1(?), then updated to 2.4.0f0. the results are the same. the erroe code is: -1074388831.
Dirk is right. I did some tests and find the max speed it can spit to the bus is 666Hz (frames/s). this is true for both channel API - sampled ouput and for frame API - CAN objects.
this will not meet the requirement of my (and many) application. I want use the NI CAN card to simulate several controllers in a vihicle to talk to one controller to be test. there are usally more than ten messages, many of them at 10ms update rate.
when using frame API network object, timestamped, the max throughput seen is 830Hz. when use "immedaite", the max throughput is ~3.15kHz (76% bus load). this rate probably meets most application,
what's Interesting is, when use channel API, signle write (sample rate 0) in a loop, use the loop to control the rate, it is able to reach same ~3.2kHz rate, although the timing is not very accuate (but the channle API is not well timed either at higher rate: 6 messages sample rate at 100Hz, ouput recent: varying from 96 to 104Hz).
the driver needs some improvement.

Concurrent modification exception inside fully synchronized code

Hello,
We have a private list
private LinkedList<ConnectionChannel> channels = new LinkedList<ConnectionChannel>();that is accessed by multiple threads. One is traversing the items and operating on them, while another thread is adding and removing items to the list. Each such access to the list is protected with
synchronized(channels) {
} scopes so that no simultanious access to this list is possible.
We do however get Concurrent Modification excpetions on this list when we do performance tests with heavy load on the sun server + heavy load on the java server itself. I cannot understand this problem in any other way that this must be a problem with the virtual machine! I know you guys probably will object to this furiously, but I am 100% sure that there is no logical error in my code.
We're running on jre 1.5..0_03. uname -a on my solaris host gives me:
SunOS trinity 5.8 Generic_108528-27 sun4u sparc SUNW,Sun-Fire-480R.
Best regards,
Helge Fredriksen

I'm very sure about this, yes. Here's all blocks where the list is used:
        synchronized (channels) {
            logger.info("Connecting observer channel from client at " + token.getSource());
            channels.add(new ConnectionChannel(token, globalObserver));
       synchronized (channels) {
            if (logger.isDebugEnabled()) {
                logger.debug("Sending local event to processing in " + channels.size() + " parallel channel(s)");
            for (ConnectionChannel connectionChanel : channels) {
                connectionChanel.send(event);
        synchronized (channels) {
            for (Iterator<ConnectionChannel> iterator = channels.iterator(); iterator.hasNext();) {
                ConnectionChannel connectionChannel = iterator.next();
                if (connectionChannel.getToken().getId().equals(token.getId())) {
                    iterator.remove();
                    logger.info("Removing connection channel to " + token.getSource());
                    return;
        }and
        synchronized (channels) {
            for (ConnectionChannel channel : channels) {
                channel.flush(5000);
        }

How do I use the buffered counting mode at a fixed frequency?

Hi-
   I'm using a PCI-6259 M-series board with the Nidaq MX
drivers in Labview 7. I am trying to use one counter to do
buffered edge counting (eg. count how many pulses appear on one input
in 400 successive time bins of 10 uS each following a digital start
pulse). That is, my inputs are:
     Start Pulse (from experiment)
     Count Pulse (from a photomultiplier tube in the experiment)
And I want to know:
    # of pulses from 0 to 10 uS after the Start Pulse
    # of pulses from 11 to 20 uS after the Start
Pulse (or, equally good, # of pulses from 0 to 20 uS....I can
subtract later)
    # of pulses from 3991 to 4000 uS after the Start Pulse (or, # of pulses from 0 to 4000 uS...same thing)
   The Count Digital Events-Buffered-Finite-Ext Clk.vi sample
appears to do half of this. I can set this up CTR0 with the Count
pulse (and possibly add the start pulse as an Arm Start....I can't use
a Start Trigger, right?). However, I need to generate a Sample
Clock Source at 100 kHz (to trigger the card to buffer the counter
value and start counting in the next bin). So, I tried to set up
CTR1 along the lines of Gen Dig Pulse Train-Finite-Dig Start.vi for
generating a finite pulse train starting on a digital trigger, and
connecting the output from CTR1 to the Sample Clock Source on
CTR0. However, I get an error -50103 saying the specified
resource is reserved if I do both at the same time. But, I can't
see any resource conflicts...the pulse generation on CTR1 works fine
alone, as well as the buffered counting on CTR0, and all the PFI pins
are different. Is there some reason I can't use both counters at
the same time?
I think I can use the FREQOUT pin on the card to generate a Sample
Clock Source at 100 kHz, since I think this is independent of CTR0
& CTR1. However, I can't trigger the FREQOUT to always start
when I get a Start Pulse (as I can if I trigger a Digital Pulse Train
to start on a digital trigger...or can I?)...so my bins will move
randomly by up to 10 uS.
   This is an unrelated topic, but is there a discussion of
the relationship between the terminology in the manuals describing the
cards (SOURCE, GATE, OUTPUT terminals) and the terminology in NidaqMX
(Source Clock/SrcClk.Source, CI.CountEdges.Term, CO.Pulse.Term)?
Eg. is CTR0.GATE always the same thing as SrcClk.Source, or does it
depend on the mode of operation? If the M-series hardware manual
says to connect something to the SOURCE input, how do I assign an
alternate PFI pin to that SOURCE input in Labview? Does it depend
on the counter type, or is it always the same?

Dave,
Hi, you brought up several questions / issues -- let me see if I can help with some of them:
   I can set this up CTR0 with the Count pulse (and possibly add the start pulse as an Arm Start....I can't use a Start Trigger, right?).
Yes, you could set up this way. Also, as far as I know you're also correct that you need to configure for an "Arm Start" trigger using the DAQmx Trigger property node. The "Arm Start" trigger is used for counter input (measurement) apps while the regular "Start Trigger" can be used for counter output (pulse generation) tasks. I don't think I've experimented with recent versions of DAQmx though so it may have changed in 7.4 or 7.5
...I tried to set up CTR1 along the lines of Gen Dig Pulse Train-Finite-Dig Start.vi for generating a finite pulse train starting on a digital trigger, and connecting the output from CTR1 to the Sample Clock Source on CTR0. However, I get an error -50103 saying the specified resource is reserved if I do both at the same time
I highlighted the problem -- the FINITE pulse train. DAQmx uses CTR0 as a "helper" when you generate a finite pulse train on CTR1. It would generate a single pulse whose width corresponds to the exact amount of time needed for CTR1 to generate its specified # of pulses.
For your specific app, I think you could generate a triggered continuous pulse train with CTR1 -- this wouldn't need to use CTR0 as a helper.  The Start Pulse would arm CTR0 at the same instant that CTR1 is started.   If you set up CTR0 to acquire on the trailing edge of CTR1's pulses, then you'll get the time bins you want.
A final slight mod would be to setup CTR0 for measuring buffered periods (set units == "Ticks") instead of counting edges. In that mode, you wouldn't have to do the subtraction at the end.
   This is an unrelated topic, but is there a discussion of the relationship between the terminology in the manuals describing the cards (SOURCE, GATE, OUTPUT terminals) and the terminology in NidaqMX (Source Clock/SrcClk.Source, CI.CountEdges.Term, CO.Pulse.Term)? Eg. is CTR0.GATE always the same thing as SrcClk.Source, or does it depend on the mode of operation?
There's an NI app note and some discussion forum hits if you search the site for "daqmx terminology."
If the M-series hardware manual says to connect something to the SOURCE input, how do I assign an alternate PFI pin to that SOURCE input in Labview? Does it depend on the counter type, or is it always the same?
Usually, that choice would be available under the DAQmx Channel property node.   There'll be some place to define where the input signal is coming from, generally with "Term" or "Terminal" as part of its name.   Sorry I can't be more specific as I'm not at my LV computer now.
Happy counting!
-Kevin P.

Multiple DAQ Cards, Multiple While Loops

I'm writing a code to support 3 small test stands that each have the same outputs and inputs as the others. I'm running them all off of an 8-slot DAQ. Each test stand has its own dedicated RTD Card and Analog Out card, but they all share a 32-channel Analog Input card.
I'd like to decouple the 3 test stands as much as possible - if I have to take out an RTD card or Analog Output card, I don't want the entire LabVIEW program to stop running. I want to be able to keep running and logging data on the other two test stands if I were working on the other test stand. Is there a way to do this with only 1 cDAQ? Is it simply a matter of creating separate, independent while loops for each test stand? If so, would LabVIEW have issues with each while loop writing its own measurement file in parallel with the others?
Thank You.

Is it safe to assume you've been running the three tests? Some of your post makes it sound like you've run them coupled while other parts make it sound like you haven't successfully had all three tests running simultaneously.
If you aren't running them, are you using hardware timing (counters) for your tasks? If you're looking to use counters, you're only using one per task, correct? The cDAQ will only have four counters total to be shared between your loops.
I can't remember off the top of my head if it is the task or the computer that reserves a cDAQ chassis. I want to say it's the computer so you can run multiple tasks from the same cDAQ. But, you'd still need to choose ctr0, ctr1, ctr2, etc in each task so that you don't share these resources between tasks.
Are you looking to pull the shared module out frequently? If not, are you gaining enough functionality to justify buying two redundant modules to gain decoupling?

PCI 6221Square signal generator

Hi all,
I am trying to generate two independent square signal with the PCI 6221, following the code example in the SDK i got how to output one signal from "ctr0" to the output pinout by means of the next code:
DAQmxCreateTask("",&taskHandle);
DAQmxCreateCOPulseChanFreq(taskHandle,"Dev3/ctr0","",DAQmx_Val_Hz,DAQmx_Val_Low,0.0,Frequency,DutyCycle);
DAQmxCfgImplicitTiming(taskHandle,DAQmx_Val_ContSamps,1000);
DAQmxStartTask(taskHandle);
My questions are:
- How can i output the square wave from ctr0 or ctr1 to one specifical channel (not the default one)?
- How can i output a second signal to another one?
- How many clocks does the 6221 have?
PD: I am using the SCB-68.
Looking forward to hearing from you.
Thanks in advance
Mensaje editado por jrojfer

Hi,
From the NIDAQmx C reference help:
The following table lists the output terminals for counter output. You can use a different PFI line for any of the output terminals.
Ctr0
Ctr1
PFI 12
PFI 13
So from that you can select any of the PFI lines....the default ones of course are Ctr0 output and Ctr1 output...
Teorically the counter clocks you talk about you can use the 20 Mhz and the 100 Khz reference clocks for generating your output signal... so precission should be enough...
Let me take a look for something more about using two counters with labwindows/cvi...
Regards,
Jaime Cabrera
NI Applications Engineering Spain

Simultaneous audio from outputs 1-2 and outputs 3-4?

Hi guys,
How do I get simultaneous audio from outputs 1-2 and outputs 3-4? I'm running Logicpro 7.1 through an M-Audio FW 410 hardware interface.
I'm trying to set things up so I can monitor through my mixing board to eleminate the weird echo sound you get in your headphones when recording.
I thought I could do it in logic by creating an ouput 1-2 and output 3-4 channel strip where ouput 3-4 will have the "i/o" plugin and in the plugin settings, Out would be set to 3-4 and input set to 1-2, this way sound will always come out of output 1-2 which will go to my speakers and output 3-4 to my mixing board. Please help!

I'm trying to set things up so I can monitor through my mixing board to eleminate the weird echo sound you get in your headphones when recording.
The 'weird echo sound' is because you are hearing Logic's software monitoring alongside direct monitoring of the inputs (either via the interface, or your mixer, depending where the headphones are plugged in.)
You can avoid this by setting the monitoring level on the record-enabled track to zero. (recording tracks have two level settings. Normal playback and record input monitoring. See page 282 in the Logic Pro reference manual.) It will not affect your record levels, or your input metering.

Async/Sync Operation with SOAP?

I have created a service interface with two operations. An Asyn iDoc operation and a Sync RFC operation. I've created a web service which the partner references. If I configure the SOAP commsn channel as Best Effort (Synchronous) then the iDoc Operation returns an error 'Only asynchronous processing supported for IDoc adapter outbound processg' BUT the RFC operation works. If I configure the web service as Exactly Once (Async) then the iDoc operation works perfectly but the RFC operation doens't return a response, it becomes Async.
Any idea how I can have an Asyn and Sync Operation in the same Service Interface, Using the same SOAP comms channel?

I do not want to tell something wrong. I have not much experience with that topic.
maybe there is something new in PI 7.11 that I do not know.
In PI 7.0 the SOAP adapter determines the QoS from channel config or from URL parameters. The SOAP adapter does not read repository and so does not know, if an interface is sync or async.
This is different, when you use WS-RM adapter with ABAP/Java proxies. Here you can use the same interface for sync and async operations.

Is it possible to play multiple videos in sync?

I need to build a website which shows multiple instruments in a band, being played in sync. Each camera/video block is a different instrument, and played together, they constitute an entire band. I have been searching high and low for ways to do this, and it definitely appears that HTML5 is out. It looks like standard FLV progressive download is out, too, because it's impossible to get 4-8 videos to sync perfectly.
I see on the FMS page that it has some abilities in synching. How would one (conceptually) go about making a webpage where say 4-8 videos begin to buffer, then when ready, play in sync? And they would also need to, when the location is changed, all sync to that location and rebuffer.
The objective is so that the user can mute/solo each video's audio, to remove an instrument(s) from the "mix" at any time.
An alternative way of doing this, I think, would be to just put the 4-8 videos together in video editing, and encode this into one video, but have each video's audio track be a separate audio channel. If FMS can do multichannel audio on one video, this might work, if it could mix them in/out individually, as the user requires.
Any insights on this problem would be greatly, greatly appreciated.
Thanks!

sandyaudiovisual,
Very good point. I was, in a way, asking both.
Yes, it appears that actually streaming multiple streams of video while visibly playing in synchronization simultaneously will be very difficult.
So, I was thinking that as an alternative, I could pre-mix the multiple "streams" of video into just one video. E.g., a single video file, with for example five squares each with a different video inside of it. So to the view, in a sense, it looks like five streaming videos in sync.
What I would then want to do, in that situation, is be able to mute/solo an audio track associated with each "video". In reality, it would be a single video channel, with five audio channels, switchable while maintaining synchronicity. Though the extra wiggle is that rather than simply switching between 1 and 5, it would be preferable to actually mix in/out channels 1-5, so that at any time, 1, 2, 3, 4, or 5 audio channels would be playing, in synchronicity, and could be muted/soloed.
To give a visual idea, there would likely be a mute and solo button composited on top of each "video" box inside the main video.
It would be easier if we could actually stream five separate video files and have them be in sync, with all stopping to wait should one need to rebuffer, and all sharing the same transport controls, because then if we want to swap out one video, we just swap that file, rather than remake the entire file in Final Cut or whatever. But this is sounding less and less possible.

Is it possible to play continous videos in my Ipad3

I want to play continous videos on my ipad....how it can be possible.

sandyaudiovisual,
Very good point. I was, in a way, asking both.
Yes, it appears that actually streaming multiple streams of video while visibly playing in synchronization simultaneously will be very difficult.
So, I was thinking that as an alternative, I could pre-mix the multiple "streams" of video into just one video. E.g., a single video file, with for example five squares each with a different video inside of it. So to the view, in a sense, it looks like five streaming videos in sync.
What I would then want to do, in that situation, is be able to mute/solo an audio track associated with each "video". In reality, it would be a single video channel, with five audio channels, switchable while maintaining synchronicity. Though the extra wiggle is that rather than simply switching between 1 and 5, it would be preferable to actually mix in/out channels 1-5, so that at any time, 1, 2, 3, 4, or 5 audio channels would be playing, in synchronicity, and could be muted/soloed.
To give a visual idea, there would likely be a mute and solo button composited on top of each "video" box inside the main video.
It would be easier if we could actually stream five separate video files and have them be in sync, with all stopping to wait should one need to rebuffer, and all sharing the same transport controls, because then if we want to swap out one video, we just swap that file, rather than remake the entire file in Final Cut or whatever. But this is sounding less and less possible.

Save multiple data independent

Hello,
Can you tell me if it is possible to save data from multiple channels, each channel with a different save trigger?
I have a cycIing machine and I need to record data from 6 channels, (that are not synchronized - different part of cycling for each channel), and to save data just when the data passed a certain value, for an among of time.
Please tell me if it is possible in Diadem, and what blocks should I use.
Thank you for the support.

Hello Marse,
Yes, that is possible. The attached example shows a data acquisition with simulated signals and two different clock rates. Saving data starts at a certain slope level for a certain number of values. But this is just one possible way to solve your request.
Greetings
Walter
Attachments:
Test.DAC ‏81 KB

Max Time Out of JDBC adapter?

Hi Experts,
In one of my scenario i need to get the acknowledgement after the records have been inserted in the database. Can i increase the timeout of JDBC channel in case of synchronous scenario ?
Or else how can i get the acknowledgement if the data to be inserted is huge (Inserting 1 million records) and takes 30 min to 1 hour time.
Thanks
Ajay Garg

Hi,
You can increase the timeout in case of synchronous scenario but is never recomended to do so. 30 minutes to 1hour is too much time.
You need to change your scenario from Sync to Async. What you can do is create a temp table wihich will update the status each time after all the rows get inserted in database.
Have another JDBC Sender chanel which will keep querying the temp table to see if the new records are inserted or not and based on that you can send response back to the Calling System.
Thanks
Amit
Reward points if suggestion is helpful

2 channel CTR0/CTR1 ouput synchronously

Similar Messages

Maybe you are looking for