Brevetti inspection machine paramaters and validation

Dear all:

Im working on validation of the inspection of ampoules in a brevetti 18/18 inspection machine. It will be a hard work, since not even the manual inspection system is validated.

Anyway, the machine is already working with product, and has serious problems in the quality output. Thanks to my previous work I have a grasp of the use of almost all of the parameters of the machine, but im not an expert.

Is there a forum, or a page or somebody to contact with info or more experience in that kind of machines? from what Ive heard (I had never talked to them), not even the Brevetti “experts” have all the knowledge about it.

Thanks in advance

Hello, what kind of help do you need?
We use the Knapp test for the calification of the Brevetti, this test can say you if the machine is working better than visual inspections.
Best regards

No, I already know about the use and application of the Knapp tests.

My problem resides on the fine tunning of the cameras parameters. Rigth now, there are problems with black particles on solution and it seems that the camera isnt detecting them. I think that due to wrong parameter use, the particles are being binarized, but also there isnt a rotating movement before the camera that inspects this kind of particle, so this may also reduce (or null) the probability of the particle being detected.

Do you have any experience on this?

Thanks.

Dear AiC,

I’m also not on expert but I have a litle experience with this.
I suppose two possible cause:

  1. as you say there is no rotating movement of particle. We had similar experience with big glass particles. The density and size are too big in this case and it settles down almost immadiately. Just laying on the bottom, the camre can not see it because there is no rotatting movement at the proper height. The best device to check the proper rotation a small checking machine which rotates your container behind a magnifing lense and lighted - simulating the operation of the inspection machine. Each manufacturer has this type of device - very useful.

The manufacturer suggested an another canera system with real picture processing, This camera investgate the bottom of the container. But the detection probability is also not too good with this very expensive solution.

  1. The brevetti machines applies lighting from below and the camera (photodiode array) detects the light refletion. If the particle is black may be the reflection is very low so the camera does not see anything. In that case better to light from background and this system detects the shadow (e.g. Eisai).

The visual inspection always has a big uncertanity - low detectability. So important to hold the reject rate at low level and to find the source of the particle.

Many thanks on your comments Asepta!!!

Yeah, I think that the main problem is there is no movement in the particle, so no mather how high it is the sensibility is, there still are black particles that are not rejected, and also, because the problematic particles arent big, but small, so the movement should “increase” the area observable by the machine.

My other concern is that why the equipment comes with a camera to detect particles with no agitation before it. Wasnt Brevetti aware of this, or the distributors just sell things without knowledge of how it works?

Anyway, what Im going to do it to adapt one of the cameras that have agitation from having light from below, to light from behind and try to detect opaque particles.

On your case of big “fixed” particles (particles that dont move or fall fast to the botttom), if you are working with a Brevetti, they have some parameters that search specifically for that kind of particles.

And yeah, tthe correct solution to apply would be not to increase rejection rate, but to decrease defectives, but sadly due to bad habits, thats not the actual concern of the company. :frowning:

Well, still, if there is somebody with expert knowledge on brevetti machines, it would be a great help if he can share it.

So glad to see we all have the same difficulties with inspection machine validation !! :confused:
Same pb with detection of small black particles… very low detection rate, even if we tried having light from below.
Do u know if there are any other way of improving the detection on that kind of particle ?
I’ve heard of a negative filter.
Have u ever heard of that ?

thanks in advance

I wouldt use light from below to detect a black particle, since it could be fainted by the glow in a white background. The best way to raise the chances that it can be detected is by incrementating its movement or allowing the camera to see it more clearly.

Thats why someone told you about filters. I dont know about “negative” filters, but there are cross polarized filters. With those, you put one in the source of light and another one after the inspectd object. Since the filters are polarized in a perpendicular way (crossed) there is no light after the filter and you see everything dark. When there is something between those filters that modifies the path of light (things with different refraction index) they “shine” because they produce light with different polarization.

I dont know if this works in a brevetti machine. As far as I know, before I worked here, some technician come and tried to install that kind of devise, but didnt work. I think that this system is better suited to particles already floating, since agitation could produce more “noise” in the image.

[quote=AiC]Dear all:

Im working on validation of the inspection of ampoules in a brevetti 18/18 inspection machine. It will be a hard work, since not even the manual inspection system is validated.

Anyway, the machine is already working with product, and has serious problems in the quality output. Thanks to my previous work I have a grasp of the use of almost all of the parameters of the machine, but im not an expert.

Is there a forum, or a page or somebody to contact with info or more experience in that kind of machines? from what Ive heard (I had never talked to them), not even the Brevetti “experts” have all the knowledge about it.

Thanks in advance[/quote]
Just entered this forum and pleasantly surprised by the commonality to a number of issues. I currently validate 2 of the older Brevetti inspection machines ( ATM 18 + ATM 32) - we validate by groupings determined by format size , colour , viscosity etc … which means we have completely different parameters for our portfolio. Like others we use the Knapp Kushner methodology. Black particles are a major issue to which binarised light settings are key , however despite intense work it is impossible to establish fool proof detection as their minimal light reflectivity aligned with their habit of suspending around the meniscus always causes issues . Brevetti as a whole struggle with technical / user friendly info , most of our evolvements have occurred through trial and error.
I have also found that employing different settings at each camera stage can have a positive impact , for example by the second camera foreign matter can be more mobile which can allow permutation at the stage to have a focus more on fibre detection which is never as proficient as glass detection. We have also modified our level detection camera with white lighting from the back which is proving as effective as could be expected for detecting black particles. I think it is key for site personnel to understand the intracities of black particle detection ( often just charred solution) and focus more on prevention upstream.