Ehrlund EHR-1 microphone... a modding project.

[abbey road d enfer]

Maybe the perceived lack of bottom is a result of characteristics of the capsule, and then especially relative to its raised frequency curve from 2000 and up?

It is a possibility; however, this type of "presence" hump is quite common in most condenser mics. here it seems to be limited to about 4dB; many cheap condenser mics have  even larger humps and tend to sound brittle. 4dB is typical of many good quality mics that don't sound too harsh or brittle.

And maybe the capsules character is exaggerated when it's combined with this actual tube circuit in the mic?

I can't really imagine the tube circuit exaggerating the phenomenon (except what I mentioned previously about the increased tension).

When googling I found the below frequency chart in a "degree thesis" concerning the development of a non tube version of the EHR-1. I don,t know for sure if the frequency chart is done with the tube circuit or the circuit under evaluation (my guess is its done with the "phantom" version under development..).

It is almost impossible to conclude anything from these graphs, since we don't know the real accuracy of the test rig. The fact that they are produced with CLIO makes me suspect they are produced in a less than ideal acoustic environment - which doesn't mean they are useless, if they make comparative measurements with calibrated mics.
The apparent dip at ca. 250Hz is a tad puzzling; I'm not sure it can be attributed to the capsule...

 

[tonedude]

I've done a simulation and I find that the grid voltage may be quite higher than I thought, at about 100V. This needs to be checked (almost impossible to check with standard instruments, but checking the cathode voltage is close enough).
If it is really the case, it is quite possible that the diaphragm is under too much tension (due to electrostatic attraction). That may explain the lack of LF response.

Abbey Road,

Do you se any problems with the lowering the 120k as a solution for the high bias voltage, or is there another way to lower the biasvoltage into the capsules  "comfort zone" (adding a plate resistor... ?)?

I measured cathode voltage to be something like 83V.

 

[abbey road d enfer]

I've done a simulation and I find that the grid voltage may be quite higher than I thought, at about 100V. This needs to be checked (almost impossible to check with standard instruments, but checking the cathode voltage is close enough).
If it is really the case, it is quite possible that the diaphragm is under too much tension (due to electrostatic attraction). That may explain the lack of LF response.

Abbey Road,

Do you se any problems with the lowering the 120k as a solution for the high bias voltage, or is there another way to lower the biasvoltage into the capsules  "comfort zone" (adding a plate resistor... ?)?

I measured cathode voltage to be something like 83V.

You can change the 120k for a lower value (about 80k seems to be adequate) or increase the 820r to about 1.5k.
BTW the bias voltage is largely tube-dependant, so replacing the 12ax with a 12au would increase the bias voltage to about 120V!.

 

[tonedude]

Abbey Road,

Ok, thank you!

I'll post back as the project moves forward (hopefully not backwards... ). Waiting for new tubes but I'll try doing the changes tomorrow.

 

[ricardo]

It is almost impossible to conclude anything from these graphs, since we don't know the real accuracy of the test rig. The fact that they are produced with CLIO makes me suspect they are produced in a less than ideal acoustic environment - which doesn't mean they are useless, if they make comparative measurements with calibrated mics.
The apparent dip at ca. 250Hz is a tad puzzling; I'm not sure it can be attributed to the capsule...

Clio is an excellent piece of test gear even for R&D.  Their measurement mikes are one of the few inexpensive ones whose calibration I trust.  For the type of measurements we are likely to want, Clio is as capable as a B&K 2012 in the same environment .. in fact more capable as the latest version has Angelo Farina's log sweep method.

But the measurement shows only 100Hz resolution so is with a VERY short FFT.  In that case, the 250Hz dip (and in fact all the 'LF' shape) is almost certainly due to 'windowing' artifacts which can trip those not versed in the art.

The different appearance of the curve below & above 1kHz also suggests these have been spliced.

The rise at 100Hz suggests that the measuring distance is very close but the poor resolution precludes any sensible assessment of this or of the LF response

So I dunno what these measurements mean but their wonkiness is nothing to do with Clio ... but with the 'expertise' of whoever made them.

 

[abbey road d enfer]

It is almost impossible to conclude anything from these graphs, since we don't know the real accuracy of the test rig. The fact that they are produced with CLIO makes me suspect they are produced in a less than ideal acoustic environment - which doesn't mean they are useless, if they make comparative measurements with calibrated mics.
The apparent dip at ca. 250Hz is a tad puzzling; I'm not sure it can be attributed to the capsule...

Clio is an excellent piece of test gear even for R&D.

If you read carefully, I was not questioning CLIO itself, but the overall test procedure.

Their measurement mikes are one of the few inexpensive ones whose calibration I trust.

[hijack] I must say I'm somewhat prejudiced against CLIO, due to the temperamental character of Audiomatica's founder and chief scientist. We used CLIO from 2000 to 2004 and we could never rely on the calibration and signor Bigi could never give us the proper procedure for the damn software to memorise its settings. Surprizingly, signor Bigi getting angry, shouting and calling us idiots didn't help... Last year we were shopping for a new test system, Audiomatica didn't agree on our request for a system on loan; we ended up buying a Klippel system. [/hijack]

But the measurement shows only 100Hz resolution so is with a VERY short FFT.  In that case, the 250Hz dip (and in fact all the 'LF' shape) is almost certainly due to 'windowing' artifacts which can trip those not versed in the art.

The different appearance of the curve below & above 1kHz also suggests these have been spliced.

The rise at 100Hz suggests that the measuring distance is very close but the poor resolution precludes any sensible assessment of this or of the LF response

So I dunno what these measurements mean but their wonkiness is nothing to do with Clio ... but with the 'expertise' of whoever made them.

The availability of relatively inexpensive test equipment is more often a curse than a blessing. In the old days, you had a $10 000 test rig with a $10 000 room treatment and a $10 000 guy who knew metrology; now you have a $1000 rig in a $1000 kitchen and a $1000 guy who thinks he's gonna show B&K how to make an acoustic measurement. Believing that windowing can cancel out all reflections in a poor acoustic environment is one of the major issues.

 

[ricardo]

[hijack][/b] I must say I'm somewhat prejudiced against CLIO, due to the temperamental character of Audiomatica's founder and chief scientist. We used CLIO from 2000 to 2004 and we could never rely on the calibration and signor Bigi could never give us the proper procedure for the damn software to memorise its settings. Surprizingly, signor Bigi getting angry, shouting and calling us idiots didn't help... Last year we were shopping for a new test system, Audiomatica didn't agree on our request for a system on loan; we ended up buying a Klippel system. [/hijack]

Well I must beg to differ on this.  I found Mauro Bigi an excellent guy to deal with and I think his technical support is fully equal to B&K, Naerum.

The availability of relatively inexpensive test equipment is more often a curse than a blessing. In the old days, you had a $10 000 test rig with a $10 000 room treatment and a $10 000 guy who knew metrology; now you have a $1000 rig in a $1000 kitchen and a $1000 guy who thinks he's gonna show B&K how to make an acoustic measurement. Believing that windowing can cancel out all reflections in a poor acoustic environment is one of the major issues.

Well, this Millenium with supa dupa computing power, I can make better measurements in my beach bum's shed than I could do with B&K 2012 and anechoic chamber last century.

But if I was to abandon the beach bum lifestyle and go back to speakers & mikes, I would want both the large anechoic AND my supa dupa stuff digital stuff.

For production testing however, I think ClioQC is THE system.  I'm biased cos I helped develop some of the ClioQC routines. 

 

[abbey road d enfer]

I found Mauro Bigi an excellent guy to deal with

Then it must be me...
Well, not really, it must be the French. The french Audiomatica distributor, who is much more patient and forgiving than me, confirmed that his dealings with il signore Bigi were sometimes heated.

 

[bockaudio]

The availability of relatively inexpensive test equipment is more often a curse than a blessing. In the old days, you had a $10 000 test rig with a $10 000 room treatment and a $10 000 guy who knew metrology; now you have a $1000 rig in a $1000 kitchen and a $1000 guy who thinks he's gonna show B&K how to make an acoustic measurement. Believing that windowing can cancel out all reflections in a poor acoustic environment is one of the major issues.

+100
add to that some of the awful sources, some with +14dB peaks right in the midband, used. Error at the source = error in result.

 

[tonedude]

A small update.

I have been tinkering around with different values for the 120k resistor in the schematic (dont know the name of this one... ), and also the cathode resistor and there is a definite improvement. Guess this is kind of like tuning a drumhead... kind of...

I tried to go as low as 60V but this also lowered the output quite a bit and the mic sounded kind of lifeless. Around 70V was mutch better because the output wasn't to low and the capsule sounded more relaxed and "at ease" compared to when the cathode voltage was around 84V. Lowering the voltage also gave the capsule the ability to produce more bottom end, relative the amount of high end coming from the capsule.

I guess the challenge is to balance the polarization voltage against the characteristics of the capsule, to find some kind of a "sweet spot".

I have installed multi turn trimmers at the above spots and I will give the circuit some more tweaking time before deciding if I better move on testing another circuit in the mic.

Fun this was!

 

[bockaudio]

How large is the membrane?

 

[tonedude]

Hi Bockaudio, I found this figure in the thesis paper about the mic and "h" is one inch, and the diameter of the green circle is 2/3 of an inch.

 

[tonedude]

US Patent:

http://www.google.com/patents/US20080205673

 

[abbey road d enfer]

US Patent:

http://www.google.com/patents/US20080205673

For me, there is nothing really new in this patent, just standard state-of-the-art. Only the shape is different, and claims that it provides better performance are not backed up by any scientific evidence. I reckon anyone could come up with the same description but pentagone or hexagone-shaped would be in the same position.

 

[tonedude]

I dont think they made any claims that there is anyting new or grounbreaking to the capsule design other than... the shape of the membrane. It's interesting to note that rectangular and triangular shaped membranes are swedish inovations. Swedes = edgy, square,,,? Dont know...  ;D


A discovery regarding EHR-1 and noise...

Yesterday I noticed that the noisefloor all of a sudden vent down another 15-20db or so (yes its really that much), just by changing to another preamp, I vent from Demeter HXM1 to my Soundskulptor MP12. The Demeter and my Metri Halo 2882 gives me aproximately the same amount of noise when using the EHR-1, a tad to much. But when using my MP12, with transformers in and out, the noisefloor is totally acceptable. Does that indicate that EHR1still has a grounding problem?

 

[abbey road d enfer]

I dont think they made any claims that there is anyting new or grounbreaking to the capsule design other than... the shape of the membrane.

Nevertheless, they applied for a patent. The essence of a patent is bringing something new and improved. If the only "newness" was the shape, it could not be granted a patent; it would just be a registered model.

Yesterday I noticed that the noisefloor all of a sudden vent down another 15-20db or so (yes its really that much), just by changing to another preamp, I vent from Demeter HXM1 to my Soundskulptor MP12. The Demeter and my Metri Halo 2882 gives me aproximately the same amount of noise when using the EHR-1, a tad to much. But when using my MP12, with transformers in and out, the noisefloor is totally acceptable. Does that indicate that EHR1still has a grounding problem?

First thing to check is electrical continuity between chassis ground, 0v and audio ground.

 

[bockaudio]

nice analysis, Ricardo.
But I still don't understand the enthusiasm for a system that is virtually unobtanium (well, perhaps not if I was Italian).

To the OP- have you considered modding it to a simpler, more conventional circuit as a baseline starting point?

 

[tonedude]

To the OP- have you considered modding it to a simpler, more conventional circuit as a baseline starting point?

Yes it might be the next step for me to do after some more tweaking with this circuit for fun. I will have to look for a circuit that fits my lundahl transformer then, I think it has quite a big ratio. There will be more questions ahead for sure...

 

[statfunk]

Just rip out the whole tube circuit, and throw away the useless China-PSU.
You can buy the new, transformerless circuit board that is used in the new Ehrlund EHR-M.
These boards cost SEK 3000 from Research Electronics in Sweden, and fits nicely inside the old EHR-1.
It works very well with the old capsule, which btw. has a broader pickup pattern (and less off-axis rejection)
due to the single membrane design and closed back plate.
The EHR-1 now sounds fantastic! Both EHR-1 and EHR-M are very sensitive, with about 12 dB higher output
than all the other mics I have. So be careful not to overload by close-miking drums, guitar-amps, sceaming lead vocals etc.
According to the designer the new circuit does not work optimally if you connect XLR input attenuators,
so beware that some types of  preamps will not be able to handle the hot signal from the microphone.

 

[jadadoog]

So here we are 10 years later and I think I have the same microphone. What did you finally do with yours?