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

[tonedude]

Just bought a cheap (200 euro) EHR-1 microphone (the one with the triangular membrane capsule) that was supposed to have trouble with noise and hum, and it sure had a problem with that. Searching and reading in on grounding schemes and PSU designs here on groupdiy helped me getting most of it fixed.

Compared to a fairly quiet mic there is still to much noise and hum coming out of it for it to be useful on solo instruments etc. But whats left I think is a result of the poor and cheap chinese PSU that comes with the mic, it's especially weak on the heater side. My plan is to rebuild the heater side of the PSU to get it as quiet as possible, and info and inspiration for that project is all over this forum... Thanx!


And now my modding question:

I also plan to change the tube in the microphone from 12ax7 to a 12au7, a Sylvania JAN 6189W, and I would appreciate some help on what to change in the circuit to adapt it to a 12au7. Increase the cathode resistor and then decrease the plate resistor? Or could I just as well swap tubes to se what i'll get out of it?

It's kind of a "weird" sounding microphone. Apparently bright sounding with a dip around 200hz, and therefore kind of weak in the lower region. But it has a very smooth and nice top that is useful and pleasing to the ear. A "shootout" comparison with a female singer: http://www.youtube.com/watch?v=IW-dgXvNni8

So, any suggestions on how to mod the circuit to get a more balanced sound out of it would also be appreciated. All kind of inspiration and/or help pointing me in the direction on what to do, or where I can read in on the subject is also appreciated.

Thanx in advance, Mattias

 

[tonedude]

Ooops, a missing resistor in the above attached schematic is corrected below.

 

[tonedude]

Gutshot...

 

[tonedude]

Some trivia about the EHR-1 microphone:

First I tried to google up info about the mic but there was not much to find. I then contacted Ehrlund Microphones trying to get some info like specifications, frequency chart and the like. But they did not want to share a thing, and there is nothing on their webpage about the EHR-1, not a single word. The thing I eventually found out was that a swedish company called JJlabs build and developed the EHR-1 using an early version of the Ehrlund triangular capsule. So the above circuit is nothing that Ehrlund Microphones is responsible for, or want to be associated with in any form... as far as I can tell.  ;D

The one mic I have bought where once on a loan to the Swedish Radio for evaluation, our national radio broadcasting network, but was returned because they felt it lacked body and bottom, and also had a problem with hum that made it unusable. After many years laying around in a music store in stockholm it was then eventually sold cheap to a guitarist and producer in stockholm as a DIY project. But he gave up on it and then selling it to me. When I took the PSU apart there was traces of someone trying to fix the hum, like extra cables drawn searching for some kind of trial and error solution to the hum, but those cables just worsened the problem. When I removed the circuit board to take a look at the solderside of it, it was obvious that three of the four filtercaps for the HT supply had their connections broken to the circuit... (Chinese manufacturing and QC working as it should, I guess  ) ...rendering the HT filter not functioning at all. So most of the hum was actually a result of a broken PSU...

The PSU in it self is a cheap generic chinese one and the grounding scheme in the mic, cable and PSU was not compatible, at least as far as I can tell after reading in on star ground in tube mics here. I then did some mods to mic, psu and cable, and after that the noise from the mic lowered substantially, but there is still some hum and noise in there that i traced to the heater circuit.

The output trafo in the mic is a unmarked Lundahl trafo. Trying to meassure it gave something like 1.5k on the primary and 20 on each secondary. It may be a custom wound transformer made specially for EHR-1, I don't know.

The mic housing for electronics and capsule is made in sweden by a mechanical workshop, and is very sturdy and nice looking if you ask me. So it's at least a nice donor body to use for another mic project if I don't get satisfied with the result after this modding adventure...  :

 

[tonedude]

Anyone?

 

[pasarski]

Don't get the circuit, never seen one like that. You're sure it's correct?

EDIT: Not true --> No polarization voltage. Has to be electret capsule then. Grid to ground 10M seems low. Would try higher value for some bottom end.

I think I'd build a whole new circuit though.

 

[tonedude]

Don't get the circuit, never seen one like that. You're sure it's correct?

I'll double check values for components and the actual circuit. But it's a really simple circuit with all the components soldered directly to the tube socket, so I haven't missed any components for sure.

No polarization voltage. Has to be electret capsule then. Grid to ground 10M seems low. Would try higher value for some bottom end.

It really lacks bottom end so i'll try that. And I read in a post in this forum that lowering the value of the coupling capacitor before transformer could actually help getting some bottom end back, so I'll do that to.

I think I'd build a whole new circuit though.

Yes, I'll probably do that. I will tinker around with the circuit a bit to see what comes out of it, and then rebuild it if I don't get satisfied with the result.


Thanx a lot,
/ Mattias

 

[tonedude]

... Grid to ground 10M seems low. Would try higher value for some bottom end.

Any suggestions on how much higher, 1G ohm, 500M ohm...?

Any problems going to high?

 

[gyraf]

I read in a post in this forum that lowering the value of the coupling capacitor before transformer could actually help getting some bottom end back, so I'll do that to.

I wouldn't count on that - try other things first.

1 - Roughly estimating your capsule at 50pF, this gives a (lowcut) cutoff frequency of 318.5Hz at 10MOhm. Here lies your bottom problem.

2 - Try with what you have of bigger resistors - 1G may work fine (or even the cheap solution - a reversed diode: this is often seen as inexpensive high-ohm resistors in mics, including eltrects). The limitation upwards in resistance is that the tube grid current must be able to return to ground through this, without offsetting tube grid DC voltage too much. Otoh, a bit of "self-bias" here can be good, if not too much.

3 - which brings us to the tube - IMHO neither ECC83 or ECC82 are anywhere near suitable for microphone front-end use, because of high miller-cap (´83) and rather high grid current (´82). In stead, try with a ECC81 or similar (12AY7, 6072) - quite often these behave much better in mics (although this use is far outside their intended specifications). If possible, stay away from new-stock ECC81's, better get a old used one if you can't find a nos.

Jakob E.

 

[tonedude]

Thanx Jakob!

Reversed diode? Is that the same as connecting the anode of the diode to the ground?

Could I place it in serial with the 10M resistor that is already there? Any problems with that?



Anyone: where do I buy these 1G resistors? Going through the white market right now, without any luck so far...


/ Mattias

Edit: Mouser is one possible source for 1 Gohm at a decent price... Banzai another one... The search is on.

 

[gyraf]

tDude,

If you post your location in your profile, help with sourcing is much easier.

And yes, the diode pointing "upwards". Using it's minute reverse leakage current as high-ohm resistor. 1N4148-like types often seen, not larger.

Jakob E.

 

[tonedude]

tDude,

If you post your location in your profile, help with sourcing is much easier.

Ok, done that. It's sweden up north in Umeå.

And yes, the diode pointing "upwards". Using it's minute reverse leakage current as high-ohm resistor. 1N4148-like types often seen, not larger.

I have some of those at home so I'll try that. Is serial with 10M ok?

 

[gyraf]

Is serial with 10M ok?

yes

RS components has high-ohmic resistors also, e.g. (100M type): http://dk.rs-online.com/web/p/hulmontering-faste-modstande/2477834/

Jakob E.

 

[tonedude]

RS components has high-ohmic resistors also, e.g. (100M type): http://dk.rs-online.com/web/p/hulmontering-faste-modstande/2477834/

Jakob E.

  Thumbs up!

 

[abbey road d enfer]

Don't get the circuit, never seen one like that. You're sure it's correct?

It is true, the circuit is rather unusual. There's a good reason for it.

No polarization voltage.

Yes there is; the 1st stage is a voltage-follower with the grid standing at about 50VDC.

Has to be electret capsule then.

No, it's an externally polarized capsule.

Grid to ground 10M seems low. Would try higher value for some bottom end.

I agree, although using the super high values needed for obtaining good LF response (>500 Meg) yields grid bias, which can be incompatible with proper operation of the tube, as Gyraf mentioned earlier.

 

[pasarski]

No polarization voltage.

Yes there is; the 1st stage is a voltage-follower with the grid standing at about 50VDC.

Oh, I see! Still don't get the circuit and the benefit of the 1st stage though. Normally you see that other way around, plate follower 1st and output from the cathode of the 2nd stage. I must be missing something obvious, have to study. These are somekind of cascaded triodes or something? Always fun to bump into a new kind of circuit in a microphone

 

[abbey road d enfer]

No polarization voltage.

Yes there is; the 1st stage is a voltage-follower with the grid standing at about 50VDC.

Oh, I see! Still don't get the circuit and the benefit of the 1st stage though. Normally you see that other way around, plate follower 1st and output from the cathode of the 2nd stage. I must be missing something obvious, have to study. These are somekind of cascaded triodes or something? Always fun to bump into a new kind of circuit in a microphone

As I'm not the designer of this circuit, I can only give an opinion.
I gave it a littl more thought and since the 10Meg resistor is bootstrapped to the cathode resistor, the actual input impedance is quite high (ca. 500Meg). Thus, I don't believe the lacking LF response comes from this implementation.
So I think the designer was concerned with the use of hard-to-find high value resistors and the difficulty of finding tubes that don't have too much grid current.
Grid current is generally not specified by manufacturers - that's one of the parameters that are tested when tubes are selected for head amp applications.
Operating the first triode as a voltage-follower has also the advantage of cancelling the Miller effect.
In a good implementation, I think the only disadvantage is the increase of intrinsic noise (3dB more than a single tube).

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.

 

[tonedude]

No polarization voltage.

Yes there is; the 1st stage is a voltage-follower with the grid standing at about 50VDC.

Oh, I see! Still don't get the circuit and the benefit of the 1st stage though. Normally you see that other way around, plate follower 1st and output from the cathode of the 2nd stage. I must be missing something obvious, have to study. These are somekind of cascaded triodes or something? Always fun to bump into a new kind of circuit in a microphone

As I'm not the designer of this circuit, I can only give an opinion.
I gave it a littl more thought and since the 10Meg resistor is bootstrapped to the cathode resistor, the actual input impedance is quite high (ca. 500Meg). Thus, I don't believe the lacking LF response comes from this implementation.
So I think the designer was concerned with the use of hard-to-find high value resistors and the difficulty of finding tubes that don't have too much grid current.
Grid current is generally not specified by manufacturers - that's one of the parameters that are tested when tubes are selected for head amp applications.
Operating the first triode as a voltage-follower has also the advantage of cancelling the Miller effect.
In a good implementation, I think the only disadvantage is the increase of intrinsic noise (3dB more than a single tube).

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.

Fantastisc, thanx for sharing your knowledge!

I will do some more testing and measuring regarding the lack of bottom end when I'm finished with the new heater supply.

I just tested it and  at least the hum is long gone (it's the heater circuit from the sony C800 PSU). Still quite a bit of noise left in there but I guess its part of the tube being a 12ax7...

 

[tonedude]

I have now done some testing and measuring and there seems to be no difference when it comes to bottom end with a reversed diode or without. I did send out a sine tone at 35, 50 and 1000 hertz through my livingroom speakers and comparing the result with my milab DC-196, placed at almost the same position, there was no significant difference between the two microphones, and between the EHR-1, with or without the reversed diode.

Measuring the cathode voltage gave me something like 83volt, so that seems to be to high if that also is the polarization voltage "driving" the capsule.

Is a possible solution just to lower the 120k resistor, or is it more to it than that?

 

[tonedude]

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

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..).