Hi there! How do you convert displacement units of mills TRUE pk-pk into mils pk-pk and mils RMS?  I am reviewing vibration settings a contractor recommended for our 2000 HP vertical pumps and notice our alarms differ from their recommendation.  We currently have them set for 9 mills RMS.  I don't know if that's too high, and I can only find severity charts for mils pk-pk and not TRUE pk-pk.


Original Post

This is interesting! 

My experience is limited so what I can say is that I am keen to learn about the reference on which your settings (9 mils rms) are based. This way of expressing vibration amplitude is not common in my small world.

I imagine this machine might be a tall low speed (possibly cooling water) pump.

Regards- Ali M. Al-Shurafa 

Luze, here are several resources for conversion of units for a specific frequency, I would assume the TRUE pk-pk is the same as pk-pk. Your RMS alarm value is .707 or 70% of peak





It depends what you're doing: 

  • Peak is the actual level of the digital samples in Logic. 
  • True Peak is the level of the analog signal that will be rebuilt if you convert Logic's digital output to analog.


Vibration severity charts conforming to ISO are based on overall vibration level in velocity mm/sec RMS. The conversion to in/sec RMS is easy. For converting from 6 mils RMS displacement to velocity in/sec RMS or (mm/sec), you have to assume that most of the vibration level is dominated by a single frequency (usually 1xSS for vertical motor-pumps). Take the pump speed (Hertz or rpm) and calculate/convert displacement to velocity to compare to the ISO severity guide. The Hydraulic Institute has velocity limits, but it generally does not apply to top (NDE) of vertical motor. There are a lot of fine points about True P-P, True Peak, Derived Peak, and RMS with much of the confusion from GE-Bently instrumentation and the differences from old analog circuits and new digital signal processing.

Online conversion calculators:

Search: convert vibration displacement to velocity



Hi, thank you all very much for responding! It was my first time posting to a blog.  After reviewing your comments, I did some additional research and found that “Peak” or “Peak-to-Peak” is a Calculated Peak not a True Peak.  GM’s vibration standard makes a reference to this on page 11 in the link below.

 The TRUE pk-to-pk value is often higher than the derived peak-to-peak value, and the conversion factors used to convert pk-to-pk into RMS cannot be used the same way for True pk-to-pk. It really depends on whether the motion is harmonic or non-harmonic. I tested this by evaluating an old displacement measurement.

 I believe the True pk-pk shows the real severity in comparison to the derived pk-pk value. Unfortunately, most severity charts and industry standards use derived mils pk-to-pk or micrometer pk-to-pk units for displacement.  I have not found severity charts that use True pk-pk units.

 I prefer to reference industry standard severity charts so staff can see where I’m getting my alarm values.  Therefore, this means I have to setup our displacement alarms in mils pk-to-pk units, or convert a severity chart to mils RMS.

 Ali, (In regards to the 9 mils RMS)

After reviewing ISO 20816-5:2018, I found that our pump’s displacement alarms should fall within 270-310 micrometers pk-pk. This equals 10-12 mils pk-pk or 3.7-4.3 mils RMS. These values are within the contractor’s recommended True pk-pk values.  Therefore, the 9 mils RMS alarm we currently have is too high.

Thank you all!

GM Vibration Standard (page 11) https://www.maintenance.org/fi...7/GMVibeStandard.pdf

 Overall Vibration Measurements and True Peak (See pages 3 and 4)  http://citeseerx.ist.psu.edu/v...ep=rep1&type=pdf



Now, I understand what you mean.

I guess despite the accuracy of the amplitude conversion from mil pp to mil rms, your way of expressing amplitudes is against the common convention. Perhaps, it could lead to confusions and mistakes.

I guess the majority of vibration practitioners and the members here in the forum would advise using mil p-p for shaft vibrations or very low-frequency structural vibrations.

On the accuracy aspect, as mentioned by others earlier,  the conversion factor (0.707) cannot hold correct if the vibration signal is complex meaning it contains two components or more. The conversion between mil pp and mil rms is true only if the waveform is a purely sinusoidal signal with a constant frequency. Maybe this is your case but I doubt that.

Regards- Ali M. Al-Shurafa


First, if someone wants a vibration specification set to mils displacement, then the sensor utilized to measure this should be measuring that as it's native units, such as a proximity probe.  That said, the proper units to read would be peak to peak.  Because displacement readings are weighted toward low frequency and higher frequencies tend to be attenuated, why would someone choose that parameter for acceptance?  I can only think of one reason for using it and that would be for assurances of clearances in turbo machinery, sleeve bearings, etc.

Velocity readings are much better suited to giving an overall look at a wide frequency range.


Ron Brook


Hi Ron,

Thanks for the feedback. The machinery we are monitoring has sleeve bearings.  They are 2000 hp pump units and operate between 450 - 695 rpms.  We are tracking displacement due to thrust (mils), vibration displacement (mils pk-pk; or in our case mils rms), and vibration velocity  (in/sec rms). I've been taught that vibration displacement is best for machines operating below 600 rpms.  I don't know how true that is, but we do monitor displacement and velocity.  For most of our smaller, higher speed pumps we use velocity readings for acceptance testing.

Kind regards,




Vibration readings can be expressed in different ways some of which are common and standard. You may want to keep using mil rms for your sleeve bearing machines but you take a risk of misinterpreting the readings if you use references based on mil p-p. Bode plot, for example, will be difficult to interpret.

Using mil rms after converting mil p-p for shaft vibration (classic horizontal machines) may twist some of the results. For example, vibration alarms are harder to trigger from rms readings when the signal has spikes. Trends in rms are smoother and less sensitive to temporary vibration changes.

Can you share some vibration plots with these mil rms readings? I'm sure it's going to be interesting to learn how you handle the analysis.

Regards- Ali M. A-Shurafa  

Add Reply