Skip to main content


I'm condidering using the relatively low cost Zoom H6 recorder for low frequency vibration logging. As I understand it the device as Auto rec functionality that I can trim to start the recording when levels are higher than a specified value... that's all good. but...

Does anybody know if the AD conversion circuitry in the Zoom H6 has a high pass filter by default, so to speak, making it unlinear below 20Hz even though the normal HP filters at 80Hz etc. are switched off. The specs say 20 to 20kHz but is that really true?

Im using a good standard accelerometer, a B&K 4507 and a BSWA MC102 signal conditioning unit for IEPE (constant current "phantom" power) so the incoming signal should be ok down to ~0.5Hz.
I do like the six channels that give 2x 3axial phase correlated recording possibility - thats why the H6 is a good candidate.

Thanks for any response at all on this issue...


pcrecord Tue, 10/07/2014 - 03:15

Humm ?? I checked on the website
I might be blind but they don't say 20 to 20kHz anywhere. I even downloaded the user manual : same result.
Anyway, the H6 was made to record music and that's what it does best.
I'm not saying it won't pickup vibrations but one thing you must realise is that a very faint signal may not trigger the record fonction.
Also, if the thresold can be low enough, it may be triggered by many things. In that regard, the manual doesn't say what are the adjustments for start and stop record.

If someone here has the unit and could answer those questions, you'll have a better idea ;)

Pejo Tue, 10/07/2014 - 04:23

Your right... it was on the hompage of the vendor... I guess they think it is a quality feature to state 20 to 20kHz for some reason.

Recording and triggering low levels is no problem, it just depends on the input Gain setting of the H6 and the accelerometer sensitivity, how many mV out per m/s^2 in out from the IEPE power unit. Auto record floors at -48dB versus ~0.7Vrms at 0dB - so that will be ok. +20dB is equivalent to 10x the acceleration value. So if I calibrate with a known acceleration at 10m/s^2 at ~0dB the acceleration level at -40dB will be 100times less = 100mm/s^2 = 0.01g. That will be small enough for triggering.

But question still stands... does anyone now if the H6 (or even other similar products) circuits have a straight frequency response curve for the recorder itself down below 20Hz or is it attenuated for low freq noise reasons. I only have used the Sound Devices 744T and that machine is even and flat down to below 0.5Hz. But it is kind of big, heavy and expensive so the H6 would seem a good idea if... I'll probably have to get one and try it out...

Thanks for reply so far...

Boswell Tue, 10/07/2014 - 05:08

I haven't done any response tests on an H6, but my H4N starts to roll off around 30Hz with the low-cut filters disabled, and I would think the two devices would be similar.

It's worth looking at [[url=http://[/URL]="http://recording.or…"]this [/]="http://recording.or…"]this [/]thread, where a researcher was looking at an H6 for recording low-frequency paranomal events.

Depending on what top end of your response curve and s/n figures you are looking for, I would say that an FM scheme is what you should be using in front of a recording device like the H6. This would give you a response up to a few KHz but amplitude and phase characteristics absolutely flat down to d.c. Signal/noise ratios of around 55-60dB should be achievable. If you demand a triggered start recording, you may have to use something like a separate amplitude comparator that generates a short high amplitude tone burst into the recorder to start the recording.

anonymous Tue, 10/07/2014 - 06:45

Pejo, post: 420012, member: 48511 wrote: I guess they think it is a quality feature to state 20 to 20kHz for some reason.

It's common these days - lots of people think that it's necessary - and some even foolishly consider it to be an actual deal breaker - if the mic (or any piece of audio gear) doesn't have those bandwidth specs.

There are many, many great mics that don't have those BW specs, and, while it's sometimes beneficial to use mics that do, it's most certainly not a "must have" feature.

For example ... the Royer R121, is a classic, well respected and much sought-after ribbon mic, one that has been used for years in pro studios world-wide; it's a wonderful mic... yet, it's bandwidth is 50hz (or so) up to around 12k, where it starts to roll of considerably. It can't - nor does it even claim to - achieve that "20/20" bandwidth... yet, it's been used in some form or another on more hit records and classic recordings than one could even begin to count.

Why? because it sounds fantastic, that's why. :)

There are many factors involved in how a mic sounds other than just the bandwidth. ;)



Pejo Tue, 10/07/2014 - 08:25

I would say that an FM scheme is what you should be using in front of a recording device like the H6.

Do you mean some sort of Frequency Modulation circuitry? So that what I record in the PCM files on the recorder is a coded representation of what the actual analog signal DC to 5kHz looked like? rather than the signal itself because that would be not trustworthy with regards to amplitude below ~30Hz. Is this correctly understood?

If so what would such a FM circuit look like? Have not worked with anything like that before ;-)

Thanks for the intrest in my problem.


Boswell Tue, 10/07/2014 - 10:43

Yes, but not really "coded" in the normal sense of the word. With FM, what you record is a parametric representation of the signal in which the instantaneous level of the original waveform is represented in the recording by frequency.

FM recording needs voltage-to-frequency converters when recording that are switched to be frequency-to-voltage conversion on replay. It's a standard instrumentation technique, used, for example, in the [="

"]Racal StoreX[/]="

"]Racal StoreX[/] series of tape recorders, but is practically unknown in the audio field. A Google search will show up papers such as [[url=http://="…"]this [/]="…"]this [/]one that use the Racal recorders in the sort of way you require. Racal instrumentation recorders are available on the second-hand market surprisingly cheaply these days, but they are very heavy, so transport costs are often more than the purchase price. Much of my day-job work is in instrumentation, so I come across FM methods quite often, although less these days, it has to be said.

Another alternative for you would be to use standard instrumentation computer recording in which you could have (say) an 8-channel interface connected via USB to a laptop. By suitable selection of sampling rate, you could get a frequency response per channel from d.c. to around 100KHz. However, these units are not particularly cheap, but may work out less than 6 channels of FM mod/demod.

MarkF48 Tue, 10/07/2014 - 14:25

Back when I was working in the electrical field, we had used some data logging Fluke DMM's to monitor and record variations in voltages and currents in circuits we had under test. These DMM's had the capability to transfer the data logs to a computer where the data could be displayed graphical. If you're going to record something as vibrations and have any meaningful data, I would tend think an instrument such as a DMM would provide this better than the Zoom recorder (although the Zoom recorder would be a fun and novel way to approach the task). The DMM would likely have better linearity over the frequency range you would be measuring and also might be easier to calibrate to the output of the accelerometer.
It's been a while since I retired, but I seem to recall we were able to set a trigger point for the data logging to start and the DMM we used could trigger off amplitude as well as some other parameters such as frequency change.
I don't recall the model we used back then and they likely have changed. One I found Googling was a... , although some cheaper models may do the job. You would need to look at the specs to see if it could do the job and the times I called a Fluke sales rep, they were quite helpful.

This might make for an interesting read also...