Hello,
I have an 07 GMC Sierra 1500 w/ 4.8L 3.23 rear axle ratio. There is a whining noise under acceleration from the rear diff. Pulling the diff cover revealed some metal chunks in the rear differential. We are going to be replacing the rear end for this reason.
I have not henceforth used the NVH microphone much. However, I did on this vehicle before replacing the rear end. The microphone on channel B is inside the vehicle. Channel A is connected to the seat frame and is only connected on the Y axis.
Here are a few questions that I have based on the capture that I have attached:
1) I assume the offending frequency is around 600-700 hz. It is a high pitch whine that you can hear slightly on the attached file around where I placed the markers. I think that is the area and, whatever is happening there, it seems to be reflected in the accelerometer capture also. Is it always the case that the accelerometer captures a sound, even when there is NO vibration felt, on the frequency view in PICO NVH?
2) Since there are different bearings inside the rear end, I suppose that we would need to know the exact construction of EACH bearing in order to do anything with this data. Is that correct?
3) Is my setup correct in this attachment? On channel A, I simply had TA259 connected on the Y axis. Should I have done some kind of vector sum reading if I was going to set it up this way without the other three channels?
4) A general question: I have seen Steve Smith set up the accelerometer to function as a microphone so that it displays DBs only when using the accelerometer. What is the setup for that? Do you simply fool the software into thinking that the accelerometer is a microphone, not an accelerometer? I know that frequencies above a certain range are not then usable. (2700 hz?).
5) A final question: what is the logic behind tires that make noise with no vibration? Since you can hear the noise, it must be a higher frequency than a T vibration normally is. What is the logic behind this? Will the tire noise appear as a harmonic of T vibrations, e.g., T100 or something?
Another question that comes to mind after reflection, which is kind of a continuation of question 3 above:
6) Is there a way to set up the three lead NVH box TA 259 so that it captures and displays (in summary format) all three axes of vibration by means of one lead hookup? For example, could I just hook up one wire into the Y axis of the TA 259, instead of all 3, and capture all three axes combined at once? In some instances, I do not want it to capture the Y axis ALONE. Do I click on a one lead interface box (TA 148) in order to achieve this? What I have in mind in particular when I ask this question is a scenario like the following: there is a bad wheel bearing, maybe from the front, but I cannot be sure which side it is. I do not feel like starting at the seat frame to identify the axis of vibration/noise. So, I hook up two accelerometers in multiple sensor mode. Then, I analyze the data via converting it to .wav files or watching the amplitude of each quadrant on the car through Time Domain to identify the offending bearing. When I set it up in this manner--and I usually always just hook it up into the Y axis alone in TA 259 when I do this--am I doing this correctly? Is the scope capturing all three axes at once, or do I stand the risk of missing something if the offending frequency is on, say, the X or Z axis?
Also:
7) A followup to question four above. I was able to use the accelerometers as microphones by fooling the software into thinking that I had a microphone hooked up. So, I guess I answered my own question here. There was an AWD Jeep that came into the shop with a slight growling/rubbing noise that you could hear as low as 5 mph. At first, it was not clear what the noise was, as there was a lot of rust buildup on the front brake caliper brackets around the points where the brake pad ears fit into them, and also very close to the rotors themselves. So, we cleaned all that up, but the noise was still present. A stethoscope revealed noise coming from a midshaft support bearing on the rear driveshaft when the vehicle was run on the rack, but how could we prove that the offending frequency heard inside the car was the same as the noise heard from the driveshaft? Since the list cost on the driveshaft was over 1500 dollars and we were dealing with an extended warranty company, I wanted to be sure. So, I set up in the following way:
1) Channels A and B are accelerometers on the front brake caliper bolts.
2) Channel C is an accelerometer on the metal bracket for the offending midshaft bearing.
3) Channel D is a true microphone inside the car.
All channels are set up as microphones and thus DB is displayed on the graph, rather than MGs. With this data, I believe that I was able to tie together the offending frequency from channel D with the offending frequency on channel C, thus allowing for pinpoint diagnostics. Playing back the audio from each channel is also informative. Although you can hear some kind of noise from channels A and B, it does not sound like the low pitch noise that is heard on channels C and D. I have attached a capture of this under "Jeepgrowl."
Thank you for your post. There is a lot to follow up on here and both Steve and I are out of the office. To ensure we give this the attention it needs it may be a few days before we can follow up. I hope this is OK.
Reading your second post, it’s clear your understanding has grown greatly since the first post. I await with interest the replies you’ll get from Ben or Steve, but I can give you my thoughts on 2 sentences from your first post. And I suspect you’ll know this anyway.
On the first sentence:
“ Is it always the case that the accelerometer captures a sound, even when there is NO vibration felt, on the frequency view in PICO NVH?”
Now, possibly you used “sound” when perhaps you meant “oscillation”, because sound is simply what we hear, within audible frequencies, as a result of oscillations, at one or more frequencies ,from a physical object. So, I’d say you’re correct in saying the accelerometer captures oscillations (within its specified frequency range), but these oscillations might not necessarily be of such a magnitude or frequency that they make themselves felt through the metalwork of the car (either as sound or vibration). So, you could say all sounds result from oscillations, but not all oscillations result in human audible sound, and if we perceive that sound as annoying or abnormal, we call it noise, and if it’s hidden or swamped by other sounds, either we’re not really aware it’s there or we call it background noise. And it’s fair to say not all vibrations produce sound - we can’t hear much below around 20Hz, but you can say all sounds we hear are produced by something vibrating, although in NVH, “vibration” tends to mean something we physically feel rather than hear, although it’s often both. That’s how I see it, anyway.
On the second sentence:
“ On channel A, I simply had TA259 connected on the Y axis. Should I have done some kind of vector sum reading if I was going to set it up this way without the other three channels?”
I infer you realise you have to have all 3 channels from the TA259 connected to the ‘scope in order for the PS7 software to make a vector sum. So, on a 4-channel ‘scope, you’d only have one channel left for your mic or another, single-channel accelerometer. With an 8-channel ‘scope, I presume you can easily have 2 accelerometers and 2 interfaces hooked up to 6 channels to give you 2 full vibration analyses and still have 2 channels left for a microphone etc. (Never tried it yet.) With a 4-channel ‘scope, if I was short of channels, I would first do a test run with all 3 axes displaying, to determine whether or not one there is one axis giving a significantly higher magnitude than the others, and then, if I needed the other channels, I’d just connect the “worst” channel. But if you’ve got 2 other channels doing nothing, you’d certainly want all 3 axes connected, to make sure you have the full picture.