Dayton Audio DAEX32EP thruster vs TT25 puck

[blekenbleu]

Has anyone compared these for the same application? Despite quite different specifications,
they do not seem to feel (or sound) tremendously different.
The puck's response is specified from 20-80Hz, but is clearly audible above 800Hz.
Meanwhile, the thruster's resonance is specified as 395Hz,
which seems about right for its electrical resonance,
but acoustic/tactile resonance seems much lower, depending on to what it is attached.

Dayton's graph shows DAEX32EP response dropping 15 dB in the octave from 140 to 70Hz,
while their DAEX32U graph shows only half that drop,
while Dayton's DAEX25CT graph shows a 5dB rise.

 

[Mr Latte]

I covered some views with my first tests Here and a few posts after that in 2019.
That page has a few posts on comparisons.

Ultra / Thruster / Puck models were tested directly to the seat.
My own original focus was using them with "Game Audio" but I soon discovered how unique they were in letting us place various Simhub effects to locations on the seat and offer multichannel effects operations within a seat.

Applying them either for having units loaded with fewer effects per units, but more units being used to better separate certain sensations. Or to extend any specific layers of effects over 4/6 units at once instead of typically only 2 on a seat.

For me the Thruster was more robust in build than the Ultra and trimming down its peak Hz may suit some people, to then increase the gain with its 40W and help feel better the higher frequencies.

 

[blekenbleu]

I covered some views with my first tests Here and a few posts after that in 2019.
That page has a few posts on comparisons.

Thanks, I thought I had read and searched this forum fairly thoroughly, but somehow missed those.
Consolidated with a table of contents, such posts could be a definitive resource.

 

[blekenbleu]

Having yet to other than sit on TT25 pucks,
standalone DAEX32EP (and Visaton BS 130) evaluations are mostly complete.
A DAEX32EP gets attached by a ring of 3M VHB single-use adhesive,
which I was reluctant to deploy prior to its intended (brake pedal) application.
Evaluations were confounded by vibrations
between that exciter and reference surface,
which is 30mm polymer-clad high density fiberboard. In practical applications,
exciters and bass shakers should be evaluated
based on their excitation of attachments,
but the reference material is substantially inert, with tactile signals buried in noise.
Consequently, tactile signals were obtained from the exciting devices,
since every action implies an equal and opposite reaction.

In comparing the DAEX32EP and Visaton BS 130, an immediate observation is that
the DAEX32EP is about 30dB more sensitive.
Drive levels barely sensed from the Visaton
would have the DAEX32EP bounding off the reference surface.

Minimizing vibration artifacts between the DAEX32EP and reference surface
required that it be driven by signal levels provoking less than 1G acceleration,
which levels were possible from the Visaton only for few frequencies,
when piston pangs became issues.

Ideally, exciters and bass shakers should be evaluated using non-contact laser Doppler vibrometers, which cost around US$1000 and up.
Contact vibrometers, typically employing piezo sensors, cost US$100 and up.
These require frequent recalibration; less than careful usage is destructive.
For purely comparative purposes,
piezo contact microphone elements are disposably cheap,
requiring simple adaptation:

• epoxy pairs together with brass disks facing out
  • this halves source resistance, electrically shields and yields 2 attaching surfaces
• epoxy a mass (e.g. steel 13mm nut) to one side
  • mass loading improves low frequency response
• wire the two piezo sensors, via shielded cable, in parallel to a high impedance audio interface
  • low impedance microphone inputs are too sensitive and attenuate low piezo frequencies

Here is a plot of multiple frequency sweeps with the DAEX32EP:

• These are disappointingly inconsistent
  • the brown trace was made with much higher mass loading
  • red and violet traces were made with various finger pressures

The brown trace is most credible, since the DAEX32EP by observation has a low frequency mechanical resonance, and rising high frequency response can be expected from a mass-loaded piezo element.
However, that much mass loading at higher amplitudes is liable to
stress both piezo elements and approx. 20-year-old epoxy.

Meanwhile, here is a plot averaging 5 DAEX32EP sweeps for identical conditions:

Here are averaged sweeps from Visaton BS 130, driven 25dB harder:

.. suspended:

.. and resting on the DAEX32EP 30mm reference surface:

By observation and comparison,
the Visaton is relatively useless at tactile frequencies of interest.
Peaks in all plots @ 60Hz represent stray AC power pickup.

These plots were made using REW, AKA Room EQ Wizard, which results can be used
for generating EqualizerAPO configuration files
to smooth tactile transducer frequency responses.

 

[blekenbleu]

Dayton TT25-16 puck overall sensitivity
is more similar to that of Thruster than Visaton.
Puck construction is comparable to that of other bass shakers, in that some mass is moved internally by alternating current thru a coil interacting with a magnet, but relatively unusual in that the enclosing case is low mass polymer, so that net inertial forces from accelerating enclosed mass are more efficiently generated, particularly compared to the Visaton, which seems a relatively poor implementation of the conventional design.

Since the TT25 polymer case is relatively low mass,
piezo signals will be more strongly influenced by
how and to what it is fastened. For first sweeps, the puck assembly was allow to float on foam included in SRS' ShakeSeat cushion:

The resulting sweep response shows a dips at 25Hz and between 100 and 300Hz, which probably result from low-Q resonances between the puck assembly and foam, allowing it to move away from the piezo, rather than transferring energy into it:

For the next sweep, SRS foam was removed,
and the plywood to which the puck was screwed
was pressed firmly against the 30mm reference surface,
with a donut of 1/2" EVA foam between
as an anti-rattle gasket and to protect the reference surface
from puck mounting screw points:

Effectively constrained, the 100-300Hz dropout is mostly gone.
The 50-60Hz resonance in this case can be blamed on the reference surface.
The 100Hz dip may be a harmonic.
For these sweeps, the puck was driven with power levels only a few dB down
from those applied to the Visaton,
which made far less impression on the reference surface.

Unlike previous sweeps, these were quite consistent,
reducing the need for averaging, but
which was anyway done for consistency.
On the topic of consistency, response dips around 3.6Hz may be
an artifact of piezo sensor glued to the puck's polymer case.
Lacking a non-contact vibrometer, we may never know...

 

[blekenbleu]

Of more interest are sweep responses in actual applications..
For sensing ShakeSeat tactile response,
a piezo pair was glued between a plastic cross of small parts cabinet drawer dividers:

..yielding this response:

.. which varies, depending on seating position and how bony is your bottom,
but less than expected.
Perception matches this plot at the low end, dropping abruptly below 40Hz,
and no longer sensible below 35 Hz, even when boosted.
Output becomes more audible than tactile (at least, by my sit bone) above 160Hz,
Low frequency response would probably improve, with a trade-off in comfort,
for ShakeSeat cushions in GFRP racing seats instead of my mesh-hammock Aeron.

 

[Mr Latte]

"Note: An exciter’s frequency response and sensitivity are completely dependent on the exciter’s designated surface. Smaller, thinner materials will tend to be louder and create a mid/tweeter response. Larger, thicker materials (with multiple exciters) will be slightly quieter but result in a more full-range sound."

 

[blekenbleu]

3M wants 72 hours for VHB to fully cure:

Thruster body in this configuration may have
more effective mass than the pedal to which it is attached,
Peak tactile response sensed by me was around 44Hz.
Unloaded measurements were confounded
by brake lever audibly rattling against its stop:

"Thruster" sweeps are from piezo glued to exciter body (above)

"pedal" sweeps are from another piezo glued to brake pedal face:

Consequently, those signals would instantaneously be out of phase,
and energy at any frequency would be influenced
by pedal and exciter mass' resonances with exciter's effective spring force.
As might be expected,
the more complex brake pedal exhibits more resonances
than the relatively rigid and dense exciter body.

The most consistent plots were of the exciter piezo
with moderate brake load:

My moccasin-clad foot evidently absorbs energy
relatively well between 40-50 and 400Hz,
with a significant anti-resonance around 100Hz.
Sandwiched between that moccasin and pedal:

.. that brake pedal piezo absorbed relatively little low frequency energy.

In sum, while this exercise may not be particularly definitive about Dayton's thruster,
it clearly suggests that brake pedal tactile energy is most effective
around 40Hz, at least for my braking foot.

 

[blekenbleu]

Having acquired a pair to laterally stimulate a traction loss project,
Thrusters were bonded to a 3.5mm aluminum strap for seat frame installation:

.. which also allowed for more representative comparison with Dayton's puck.

First, since the strap could be firmly held against the 30mm reference surface,
the Thruster could be driven harder,
improving signal-to-noise and consistency of plots,
but also making more evident that
the 30mm reference surface is less inert than was supposed:

A high Q resonance previously noted near 60Hz is still evident,
but the broad peak around 600Hz shows this surface less damped than expected.
More surprising is relatively large response below 20Hz, which is not sensed.
Unobvious from this graph is that 25Hz energy injected by the Thruster
was sensed in a Logitech mouse half a meter distant.
Sympathetic vibrations and rattles were issues with the Thruster
that had not been with the puck, much less Visaton,
both driven with over 10dB more power.

Here is a plot for piezo glued to Thruster body, moving in reaction to 30mm surface:

Individual sweeps, as well as average, show how consistent are results above 28Hz.
These sweeps were made with settings identical to the above sweeps,
demonstrating how much a Thruster energized the 30mm "reference" surface
above 200Hz and 9 to 30Hz. For convenient comparison, here is puck sweep result:

Since measured with different gain settings, only curve shapes are comparable,
but note that averaged puck response varies overall more smoothly about 70dB,
while averaged Thruster response varies about 36dB,
albeit with an abrupt transition between 30 and 50Hz, which can be addressed
by equalization, given how much more tactile energy Thrusters produce.
The puck seems penalized by a relatively low moving mass
less able to transfer higher frequency energy into that reference surface
(until its broad resonance starting at 300Hz) so well as can the Thruster.
Considering responses between 100Hz and 15Hz, below which neither are sensible,
the puck is +/-8dB, compared to +/-18 for the Thruster,
which has more output to equalize away, if need be.

FWIW, checked with a multimeter on initial receipt,
Thrusters have appreciably lower DC resistance
than is typical e.g. for subwoofers.
Consequently, wiring a pair in series, rather than parallel,
seems more reasonable for low frequency tactile service
as a load for conventional amplifiers.

 

[blekenbleu]

With no relevant TT25 puck comparison, here is a dual lateral Thruster configuration:

Since this frame is on the back of a swivel chair,
Thrusters are driving into effectively no spring constant for low frequencies..
Of course, a seated human represents some distributed mass and spring.
Only one of two Thrusters has a piezo sensor attached,
so unless responses are perfectly matched,
sweep reading from one Thruster body might show cancellations from the other:

Being a relatively complex structure,
multiple frame resonances are also to be expected:

.. but overal response below 150Hz is curiously uniform,
particularly with a sensed resonance around 43Hz,
which might have been evident in an unoccupied sweep..

 

[blekenbleu]

Rather than simply replace DAEX32EP-4 burned out while fiddling with brake pedal 14Hz feedback with another of the same, I opted instead for a DAEX30HESF-4,
which specifies a slightly lower resonance.
Comparative pictures are already shown by @bassun:

Here are a couple side by sides of the Thruster vs the HESF. DAEX32EP-4 vs DAEX30HESF-4
View attachment 355170View attachment 355171View attachment 355172View attachment 355173

On inspection, lower resonance probably results from the DAEX30HESF-4 lacking the spring steel supplementary suspension of the DAEX32EP-4. FWIW, the DAEX32EP-4's finned black surround is seemingly NOT a heat sink;
it is of a substantially non-conductive polymer, adding little mass.
The DAEX30HESF-4's VHB mounting ring is relatively and concerningly small.
Whether its advertised "High Efficiency Steered Flux Exciter with Shielding" magnet circuit actually differs from that of the DAEX32EP-4 seems unlikely; magnet external dimensions are very similar...
Its more open construction may afford better voice coil cooling,
and its solder tabs for wiring attachment are distinctly more robust.

 

[blekenbleu]

DAEX30HESF-4, which specifies a slightly lower resonance

Sweeps with piezo glued to exciter magnet:

..compared to thruster:

Piezo glued to 30mm reference surface:

.. compared to Thruster:

.. so despite somewhat lower mass, an HESF more effectively
drives energy into the 30mm surface between 70 and 300Hz

 

[blekenbleu]

Adding mass (magnet scavenged from a Thruster) to the HESF increased energy driven into 30mm reference around 24Hz, but forfeited 30-40Hz energy :

Mounted nearly vertically, lateral load from added mass on HESF suspension,
which lacks Thruster's steel spring, might be problematic

Attached to the back of a G29 brake pedal,
a DAEX30HESF-4 works better than expected:

... compared to the Thruster:

.. and ignoring absolute SPLs, which were not controlled,
the Thruster has +/-20dB between 20 and 600Hz,
while the HESF is +/-12dB over the same range.

 

[miwashi]

Any update on this? I'm interested in getting a DAEX series transducer, but am not sure which one to get as there are so many variants.

 

[blekenbleu]

there are so many variants

Only 3 have 40 Watt ratings, and DAEX32QMB-4 mounting seems inconvenient:

Between the other two, DAEX30HESF-4 had more uniform frequency response in my testing and (my guess) is less liable to overheating, while the DAEX32EP-4 Thruster's spring steel suspension may be more robust against physical abuse, with potentially better long term voice coil centering stability e.g. when mounted vertically. DAEX30HESF-4 adhesive mounting ring is smaller diameter than that for DAEX32EP-4, which may be more suitable in some applications, although overall space requirements are similar.

5 Nov 2021 update:
I have since read of multiple DAEX32EP-4 spring steel suspension failures,
presumably metal fatigue and possibly from being overdriven.

 

[blekenbleu]

Contact vibrometers, typically employing piezo sensors, cost US$100 and up.
These require frequent recalibration; less than careful usage is destructive.

I recently stumbled upon an
NIST paper for calibrating piezo accelerometers by signal insertion.
The basic concept is that piezos respond to AC voltage stimulation
as well as mechanical vibration.
By stimulating a grounded piezo in series with a resistor,
driven from a signal generator.
the ratio of signal at the piezo to resistor junction
to that at the resistor to generator junction
should generally agree with more conventional calibration
by known mechanical vibration.
It should at least identify resonant frequencies for e.g. our mass-loaded piezos,
so that we avoid misplaced credibility for responses near those resonances.

 

[miwashi]

Thanks for all the replies. So you all would recommend going for 40w ones? Is it possible to use these exciters totally and not use Dayton tt25 pucks?

 

[blekenbleu]

So you all would recommend going for 40w ones?

Among exciters, reasons to go for less that 40W:

• 40W exciters are too large to fit intended locations
• unavailability of 40W exciters

Both of those are marginal cases. First, so-called 40W exciters
will overheat and fail if driven by anything near 40 Watts continuously.
Second, even 40W exciters are incapable of driving masses substantially
greater than their own over the interesting range of tactile frequencies.

Is it possible to use these exciters totally and not use Dayton tt25 pucks?

No, pucks are basically small bass shakers: they include a mass that is excited by
their electromagnetic field, so not depending on attachment to some external mass
to deliver tactile energy. This makes pucks better for applications directly
delivering tactile energy to body parts, e.g. inserted in cushions or seat upholstery.

Exciters, on the other hand, more efficiently produce tactile energy
in low masses to which they are attached, e.g. pedals.
When an exciter is attached to a mass greater than its own, it is at a disadvantage.
IMO, many folks attaching exciters to their seat shells
would get better tactile effects using the equivalent of SRS ShakeSeat or Shake Plus,
which are simply arrays of 4 or 6 Dayton pucks in thin foam cushions.

Exciters attached to larger masses will mostly generate sensible tactile effects
only at frequencies near resonances in those structures.